When I first started writing this newsletter, I had a general idea of who might find these topics interesting. But looking at our community today, I am incredibly humbled by the sheer depth of expertise represented in our subscriber list.

We have university library deans, technologists, open infrastructure organizations, repository managers, and academic researchers reading along every week.

Because this community is so uniquely specialized, I want to make sure I am honoring your inbox. I want to ensure that every essay, deep dive, and strategic critique I publish is directly relevant to the real-world challenges, decisions, and technology issues each of you navigates day to day.

To help me do that, could you take 45 seconds to answer these four quick questions? Your responses will directly shape my editorial roadmap for the rest of the year.

Start Survey

Thank you so much for being a part of this community, for sharing your insights in the comments, and for helping me make this newsletter as valuable as possible for the work you do. I read every single response, and I can’t wait to see where we take this publication next.

• Part 1: Reclaiming Information Sovereignty

• Part 2: The Forcing Function

• Part 3: The Ideological Storm

• Part 4: A Shared Reality

• Part 5: Governance and the Commons

• Part 6: Reclaiming our Seat at the Table (this post)

We began this series with my confessions of struggle, skepticism, and sheer exhaustion. But as I look at the issues facing society, I’ve realized that giving up isn’t an option.

Here are the three things I urge our profession to do in order to reclaim its own intellectual sovereignty:

1. Move into ‘The Medium Place.’ For too long, our profession has built silos that treat technologists as a support service rather than intellectual partners. I am asking you to dismantle these dynamics by ending the isolation within our own organizations. Building a digital commons requires more than just hiring IT staff; it requires us to build cultural competency in technology. We cannot simply adopt the vocabulary of computer science while ignoring what makes the technology work. Instead, approach technologists not with a support ticket, but with a seat at the table. Consider asking your systems administrator or developer to chat over coffee because you want to understand more about them, how they got on their career path, and what it is that they work on. Engage them. Learn their language with the same rigor you apply to your own. If you don’t understand a concept, do the hard work of researching it yourself before you ask for an explanation. Doing so fosters the mutual trust necessary for a robust digital commons.

2. Reclaim the philosophy of our values. This is the answer to my second confession—the skepticism I felt when I first heard Bohyun talk about AI. I didn’t have the foundation to judge it fairly. But after spending some time with the technology, I see I had the toolkit to understand it all along. I’ve always been frustrated that Library School teaches us how to do the work but rarely why it matters on a civilizational scale. We need to rethink the profession’s curriculum so that we read the great philosophers who argued for understanding truth, knowledge, and epistemology—how we define what is actually true and how we know it. Even better, take the time to understand how different societies learn and preserve their cultures—I’m trying to do that right now by studying non-Western traditions of stewardship, which prioritize relational accountability and community permissions over extractive consumption. You cannot architect a new knowledge infrastructure if you haven’t studied the foundations of knowledge itself.

3. Embrace AI to get to Human Reclamation. Finally, I’ve realized that I cannot keep living under the old rules because the world has already changed. To survive the demographic storm and the loss of institutional knowledge, I’ve embraced AI to reclaim my own humanity. I wrote the majority of this keynote using voice typing; I use Speechify and News Over Audio to ‘read’ while walking my dog; I use the Zoom AI Companion to capture the endless to-do lists I am too overwhelmed to write down myself. All of this has been made possible by AI and serves as a personal lifeboat, carving out the space needed to remain present for my family and my community. But while AI may be my personal lifeboat, we as a profession must use this reclaimed capacity to build the collective shore. And we must do this with our eyes wide open. I am acutely aware of the paradox here: I am relying on a lifeboat built by the very forces threatening to flood our ecosystem. But we must lean into them because, as Sun Tzu famously advised, we must ‘know thy enemy.’

Within our profession, there is a dangerous tendency to ignore AI out of a sense of moral purity, but we are being left behind in the process. So I’m advocating for the profession to develop a deeper, more practical understanding of these systems. We cannot architect a future that protects our values if we don’t understand the machinery currently destroying them. If we are to heed Hannah Arendt’s warning and prevent our institutions from becoming an ‘infrastructure of denial,’ we must set up an account and play with the tools so that we can be the auditors of the black-box technology corporations are creating.

I encourage you to use AI to automate a manual task—whether that is asking an LLM to proofread your writing or having it summarize a dense 50-page PDF you already read but can’t quite remember (or maybe that’s just me?). Only by using these fallible tools can we develop the empathy to understand why our users are succumbing to them. But more importantly, through using them, we will gain the expertise needed to lead our communities back to a shared reality.

Last month, Cecilia Preston reminded me that in 1876, the founders of the ALA met in Philadelphia to solve “the educational problems of the country.” They weren’t just minding the books; they were attempting to architect the infrastructure a democracy needs to survive. Today, the “Frontier Labs” are trying to architect a world of engagement and profit, all while pretending it is a new infrastructure for knowledge. Don’t let them do that.

Your role—our role—is to provide an alternative AI infrastructure. Using Ostrom’s principles, we can come together to define what open means in this new age. We need to define clear boundaries around the content we steward to protect it from bad-faith harvesting. We need to help communities identify how they want their content used on the open internet. And we need to build nested enterprises that link local curation to a global, non-commercial network that we trust. The intent of these big technology corporations is to capture our attention and never let it go; our intent is to nurture the independent mind so that it can improve society. Like in 1876, we are once again the heroes democracy needs to survive.

I am asking you, the members of our profession to do this work so that we can reclaim our seat at the table. Not as a group of people minding the books, but as the architects of information infrastructure that belongs to the people.

It’s time to start building.

• Part 1: Reclaiming Information Sovereignty

• Part 2: The Forcing Function

• Part 3: The Ideological Storm

• Part 4: A Shared Reality

• Part 5: Governance and the Commons (this post)

• Part 6: Reclaiming our Seat at the Table

In preparing for this keynote, I thought deeply about what differentiates galleries, libraries, archives, and museums (GLAMs) from big technology corporations. And what I realized is all that really matters is the obvious: GLAMs are focused on cultivating, sustaining, and reinforcing the communities they serve. Big tech corporations are focused on acquiring and profiting off consumers.

If that is the major difference, what structures must we build to ensure our communities can continue to prioritize the values that will help them thrive now and into the future?

For too long, we’ve been told there are only two ways to manage a shared resource:

1. Big Government Regulation: Top-down state control that regulates the resource into strict, bureaucratic submission.

2. Corporate Privatization: Handing the resource over to private markets to own, manage, gatekeep, and ultimately lease or sell back to us.

This is the way Locke, Hobbes, and other philosophers have described the world for hundreds of years.

But there is a third way.

Elinor Ostrom, the first woman to win the Nobel Prize in Economics, spent her life studying the Commons. She defined the Commons as a resource managed by the community that depends on it. More specifically, she examined how people successfully shared fisheries, forests, and irrigation systems without these systems collapsing into greed or neglect.

To understand how Ostrom rewrote the rules of economics, we first need to unpack the baseline vocabulary used to classify resources. Standard economics categorizes all goods based on two fundamental variables:

1. Excludability: How easy or costly is it to prevent someone from accessing or using the resource? High excludability means you can easily keep people out (e.g., a locked building, a paywalled website). While low excludability means it is incredibly difficult or practically impossible to prevent access (e.g., the open ocean, the public internet).

2. Rivalry: Does one person’s consumption of the resource reduce the amount or quality available to others? Highly rivalrous means consumption is zero-sum (e.g., if I eat an apple, you cannot eat that same apple). While non-rivalrous means simultaneous use is infinite (e.g., my listening to a radio broadcast doesn’t diminish your ability to listen to it).

A Common-Pool Resource (CPR) is the trickiest of them all: it is low-excludability but highly rivalrous. It is a shared ecosystem where anyone can use the resource, but every time you use it, it leaves less for everyone else. This is where the risk of exploitation is highest—and it is precisely the ecosystem Ostrom set out to define.

Did something I said spark an idea?

Leave a comment

In her seminal book, Governing the Commons, Ostrom outlined eight design principles that serve as the foundation for building trust and sustainability in shared spaces:

1. Clear Boundaries: Knowing who is “in” and who is “out” of the user group, and what constitutes the shared resource.

2. Local Congruence: Rules must fit the actual local context, culture, and biophysical environment.

3. Collective-Choice Arrangements: The people affected by the rules must have a seat at the table when those rules are written or changed.

4. Accountable Monitoring: The community itself must monitor behavior, ensuring accountability.

5. Graduated Sanctions: Rule-breakers face escalating penalties—starting as gentle, constructive reminders before moving to harsher consequences.

6. Conflict-Resolution Mechanisms: Quick, low-cost, local arenas to resolve disputes.

7. Self-Determination: External authorities must recognize the community’s right to self-govern.

8. Nested Enterprises: For larger systems, nested layers of governance are built from the local level up to the global level.

For Ostrom, a commons is not merely a free-for-all pile of resources. It is a living, breathing social system rooted in a web of mutual relationships. Governance, in her framework, is ultimately a mechanism for constructing and maintaining trust.

Toward the end of her career, alongside Charlotte Hess, Ostrom realized her work applied directly to the scholarly record, not just to the initial fisheries and grazing lands she had originally discussed. Together, they pioneered the concept of the Knowledge Commons. Within this framework, we can view knowledge as a distinct type of CPR.

In classical economics, digital knowledge is viewed as “non-rivalrous”—if I read an open-access article, it doesn’t prevent you from reading it too. But Ostrom and Hess recognized a critical nuance: while information itself may be non-rivalrous, the access and maintenance infrastructure supporting it is.

Which brings us to my favorite topic: AI web scraping.

If you have ideas, please reach out. I’d love to hear what you are thinking or what you’d like me to write about next.

When AI companies unleash aggressive, unthrottled web-crawling bots to scrape the content on GLAM websites, they aren’t just reading (which, remember, we said is non-rivalrous). Instead, they are actively consuming our infrastructure’s bandwidth, degrading server performance, increasing infrastructure costs, and shutting down human access to the content. Even worse they are:

• Misappropriating cultures by scraping sacred or sensitive knowledge, stripping away community-defined permissions.

• Decoupling knowledge from its provenance, severing the historical context, and destroying the metadata and curation work painstakingly done by GLAM professionals.

• Eroding knowledge sovereignty, turning curated digital heritage into anonymous, homogenized training tokens for commercial LLMs.

By weaponizing openness to harvest data and use it to train AI models, they turn a CPR into a playground for corporate greed.

This is precisely why Ostrom’s design principles are so vital for the survival of GLAMs in the age of artificial intelligence. They provide a blueprint for defending our digital infrastructure against corporations.

By applying her principles, we can design a resilient, active, relational ecosystem for human beings. To borrow from my own writing:

[We] need to evolve from being “open by default” to being “open with thoughtfulness.” This entails creating environments in which agreement, governance, and collaboration guide participation. Genuine openness supports the commons through shared responsibility, rather than through indiscriminate exposure.

~myself in Openness Has Limits

But as we look to architect “open with thoughtfulness”, we can’t just focus on developing flexible schemas, model agreements, or solid technologies. We have to begin by reclaiming our roots—returning to the foundational philosophies of epistemology, truth, and the process of knowledge creation. Next week, in the final installment of this series, we will dive deeper into what it means to step out of our comfort zones, honor those intellectual origins, and cultivate a modern profession capable of anchoring our communities in a changing world.

• Part 1: Reclaiming Information Sovereignty

• Part 2: The Forcing Function

• Part 3: The Ideological Storm

• Part 4: A Shared Reality

• Part 5: Governance and the Commons

• Part 6: Reclaiming our Seat at the Table

Over the last year or so, as I’ve been writing this Substack, I’ve learned that so much about information and technology is wrapped up in the politics of society. One of my biggest takeaways has been that when we lose agency over information and technology, we lose the ability to live together in a shared reality.

In 1975, Hannah Arendt gave a Bicentennial Address titled "Home to Roost." She warned that a deeper corruption of political reality occurs when greater emphasis is placed on “image-making.” This is something I discussed in my iPres keynote, and I’ll link to that post below. Essentially, Arendt believed the pursuit of image-making can hollow out the government’s moral judgment and turn institutional machinery into the infrastructure of denial. She argued that a shared reality is the only thing that sustains political life. By replacing truth with image, citizens become alienated from one another and from the shared reality they depend on. So, to put it in today’s terms, when we retreat into our algorithmic boxes, we aren’t just being distracted; we are allowing the foundation of society to rot away.

I have one of those boxes. My box is on Instagram. At present, it’s filled with fashion, dogs, cats, crochet and other crafts, and home renovations. My box is a space designed for zoning out, not learning. However, unlike the vast majority of social media users, I recognize that I am in a box. In grad school, I learned how fine-tuning algorithms can change search results; basically, I studied the mechanics of social media. Many users don’t realize that endless scrolling isn’t a choice that you make; algorithms are replacing users’ ability to search for and choose content with automated curation. Algorithms are taking away humans’ agency over the information they consume and the technology they engage with.

I’ve learned the mechanics of these tools, not to endorse them, but to reveal exactly how agency is being taken away from us. I know from reading Arendt that trust in a democracy is a calculated decision based on shared reality. When a citizen can no longer verify the shared reality, they cannot trust the knowledge they receive from anyone.

Lately, I’ve been spending a lot of time with Mark Coeckelbergh’s book, Why AI Undermines Democracy and What to Do About It. Reading large sections of his work feels so aligned with my own thoughts that I wonder if I might have written them myself. Coeckelbergh explores how contemporary AI—particularly social media and Large Language Models—confines us within the algorithmic boxes I previously mentioned. He highlights the immense power and control accumulated by AI corporations and their persistent efforts to maintain that dominance over the technologies we use.

In Chapter 6, my alignment with Coeckelbergh comes to an abrupt end. In this chapter, he discusses how we can pull AI back from corporate hegemony. He identifies a lineup of roles he believes are essential to save the public sphere: independent journalists, non-corporate editors, and skilled mediators. He even calls on computer scientists to stop viewing their work as neutral and start seeing the political weight of every line of code they write.

But I think Coeckelbergh’s focus on job titles is a little short-sighted (I mostly just kept saying, "what about librarians!"); rather than identifying particular roles to be filled, I wanted him to identify skills and abilities that society can curate in the citizenry. The goal of this approach would be to empower society with more control over its information consumption. By integrating these capabilities into our educational frameworks, we can foster a deeper collective understanding of how these technologies and the information flowing through them function.

In an effort to build upon his work, I have identified the specific skills and competencies that I believe Coeckelbergh envisions as essential components of our social fabric.

I believe Coeckelbergh is looking for a citizenry equipped to handle what he calls our modern political-epistemic challenges—specifically, by restoring our capacity for critical evaluation and information veracity through rigorous fact-checking. He is calling for an ethical and political commitment to the common good, rooted in a republican conception of democracy (that’s small-r republican for those of you from the US), alongside a fierce insistence on independence and non-corporate alignment. He identifies a vital need for independent editors and mediators who can practice a form of mediatory neutrality, acting as democratic gatekeepers to guard against anti-democratic populism and authoritarianism. And finally, he demands that our technologies be anchored in core principles like transparency, accountability, and diversity—ensuring that the digital ecosystem ultimately answers to human values.

But as I was reading this chapter, I realized something. Guess what that list is? That is a list of skills and abilities you need to fulfill a library’s mission. Those qualities allow libraries to prioritize provenance, truth, and the preservation of the human record. This list of qualities he is asking the world to embed in AI companies is one that librarians have spent centuries refining. And the same is true for computer science, data science, and information technology (which I believe are newer forms of librarianship). Now we just need to figure out how to scale these skills and abilities to the entire citizenry (easy peasy, right?).

Did something I said spark an idea?

Leave a comment

But this tension between a citizenry that needs open information to maintain agency and centralized structures that try to control it is not a new crisis manufactured by modern algorithms. It is the foundational struggle of democracy itself.

If we look at how information asymmetry has shaped human governance, we see that the free exchange of information has always been a necessary part of the formula for democracy. In his article, The Origin of Democracy, J.L. Gillin explores how the exchange of information and public debate served as the primary engine of primitive tribal democracies, where leadership relied on horizontal persuasion rather than coercive command. But as society grew, knowledge was increasingly hoarded by elites. The democratization of information through universal education became the great equalizer, breaking down silos so a diverse population could move beyond these silos to forge a shared reality.

Conversely, when information is choked off, democracy fails. Ahmed and Stasavage argue in Origins of Early Democracy that early democratic practices were actually born as a practical solution to information asymmetry—rulers shared power with local collective councils because they needed accurate local crop data to tax fairly without sparking rebellion. It was a mutually beneficial exchange. However, these early democracies began to fail the moment states built centralized bureaucracies and writing systems to measure productivity on their own, effectively stripping subjects of agency over their own information.

This historical pattern brings us right back to Coeckelbergh. In Chapter 8 of his book, he focuses on this exact threat in our modern digital ecosystem. He argues that communication should be understood not just as the transfer of data, but as the active creation of a common world and the common good. He warns that modern AI and algorithms threaten this by isolating citizens into silos, destroying the shared reality necessary for political trust. To Coeckelbergh, the ultimate role of information is to transcend individual perspectives, which requires an information commons completely free from corporate hegemony.

When we look at these pieces together, we can map them onto the classic debates of political philosophy regarding why humans form societies in the first place:

• J.L. Gillin represents the optimistic idealist (like Rousseau), believing that equal access to information naturally dissolves inequality and fosters a harmonious public mind.

• Ahmed and Stasavage adopt a cold realism (like Hobbes), framing information exchange not as a community-building ideal, but as a calculated survival mechanism born of mutual suspicion between self-interested actors.

• Mark Coeckelbergh occupies the middle ground (as Locke does), pragmatically arguing that while communication can build a democratic common world, it requires a fiercely protected, structured environment to survive.

Whether you view information exchange as a romantic ideal, a pragmatic calculation, or a tool for a shared world, the conclusion is the same: democracy requires a protected space where information resources are shared and managed by the community, not corporate overlords.

And I think if we want to find the blueprints for how to actually build and sustain the digital commons Coeckelbergh talks about, we need to turn to the work of the first woman to receive the Nobel Prize in Economics, Elinor Ostrom. We’ll unpack her blueprint next week in our fifth installment of this series.

• Part 1: Reclaiming Information Sovereignty

• Part 2: The Forcing Function

• Part 3: The Ideological Storm

• Part 4: A Shared Reality

• Part 5: Governance and the Commons

• Part 6: Reclaiming our Seat at the Table

Back in part 1 of this series, I mentioned that I wasn’t thrilled with AI, which is one of the reasons I questioned my friend Bohyun, who was an early evangelist for cultural heritage organizations.

One reason I distanced myself from AI was that the field’s prominent voices often promoted deeply disturbing ideologies. Many of their theories were rooted in Eugenics, and they held beliefs I find fundamentally reprehensible.

To give you an idea, I was already familiar with figures like Peter Thiel, a man often characterized by his techno-utopianism and a clear skepticism toward democratic ideals. My wariness extended to Elon Musk as well; initially, it was his proximity to Thiel that gave me pause, but as I dug deeper into his role at the helm of Tesla and SpaceX, his underlying moral compass became increasingly questionable to me. I had also caught the growing unrest surrounding Nick Bostrom, whose listserv correspondence revealed deeply problematic perspectives and ideas about genetic engineering, which, to me, felt like a new version of some of the ideas put forward by traditional eugenicists. It became impossible to ignore that the tycoons leading today’s major AI “Frontier Labs” (e.g., Altman, Amodei, etc.) were inextricably linked to these same individuals, and so I remain extremely skeptical.

Did something I said spark an idea?

Leave a comment

And then I stumbled across an Instagram reel—ironic, I know—that mentioned the TESCREAL bundle. A quick search led me to Dr. Timnit Gebru and Dr. Émile Torres. Their research, published in exposes the beliefs behind today’s technology-driven future.

If you don’t know, TESCREAL stands for: Transhumanism, Extropianism, Singularitarianism, Cosmism, Rationalism, Effective Altruism, and Longtermism. The philosophies are often rooted in eugenics and techno-utopianism. And I want to be clear, not all TESCREALists know about the origins of their beliefs.

When I learned about the concept of the TESCREAL bundle of ideologies, it seemed par for the course for folks in Silicon Valley. It essentially represents a worldview where technology is meant to replace the human experience rather than enhance it. In that world, we aren’t people with stories and values; we are just data points waiting to be uploaded, optimized, or replaced.

However, it’s important to remember that AI is not some sentient being; it is a mirror of its architects. Once I looked past the deeply disturbing ideologies of prominent voices and focused on the technical foundations, I was able to see that these models are nothing more than math and statistics, but more importantly, they are inherently susceptible to the biases and values embedded in the data they consume—and given the current state of the internet, that is hardly a comforting thought. These systems operate through a mechanical process called Tokenization.

Imagine every word, or even a fragment of a word, being assigned a numerical value—a “token.” In its simplest terms, tokenization is the method of decomposing human language into smaller, bite-sized components so that an AI model can ingest it. Because these models are essentially colossal mathematical engines, they possess no capacity to “read” words; they only compute numbers. Tokenization serves as the bridge between our human expression and their binary reality.

A token can represent a complete word, a lone character, or a “subword” chunk. For instance, while a familiar word like “apple” might exist as a single token, a more intricate term like “tokenization” is often broken down into fragments such as “token,” “iza,” and “tion.” Each fragment is granted a unique ID number. When you input a prompt, the system translates your text into a sequence of these identifiers, performs its mathematical heavy lifting, and then renders those numbers as human-readable text on your screen.

When you engage with an AI, it isn’t “thinking” in any human sense. It is merely scanning a sequence of numbers and calculating the statistical probability of the next number based on the billions of examples it has ingested. It isn’t magic; it is a sequence. Once we demystify this process, we reclaim the agency to audit these tools for bias, shape their governance, and deploy them to defend the scholarly values we have stewarded for generations. The pressing question we must face is: where do the billions of documents required to fuel a statistical engine like this actually come from? AI’s relentless hunger for data is what defines the next chapter of our story.

I’ve been vocal on my Substack and in various talks about the predatory nature of AI bots and their voracious appetite for content to fuel their models, but I’m not going to assume you’ve read all of that (see below, though, if you want to).

So quickly, these companies operate under the delusion that if they simply scale the amount of data they train on, they can reach the holy grail: Artificial General Intelligence (AGI). So they either send out web-scraping bots to crawl sites and capture all the data, or they rely on the Common Crawl dataset to train their models and employ content moderation and data-labeling firms to scrub and refine the datasets they use for training.

But there is a shortcut that allows them to bypass content moderation and data labeling firms entirely: us. Libraries have been the original data labelers for centuries. We are the curators and collectors of the human record. They can effortlessly crawl our repositories and use our work to curate and describe content to train their systems. And the most unsettling question remains: what happens when even that well runs dry?

More recently, they have pivoted toward the content we haven’t yet digitized. In this screenshot, you can see an invitation I received from Google, asking Emory to join the Google Books project. What’s missing from this frame is the inquiry a colleague received from Meta, specifically, a request for materials within our Special Collections. They are acutely aware that high-quality, curated content still exists beyond the reach of their current crawlers, and they know there is no repository more qualified than a library to provide it.

While we in the cultural heritage sector have long maintained a bright line between archival and non-archival materials, research data and scholarly articles, these entities do not care about such fine-grained distinctions; they only see content. The moment we lose the ability to steward and control the intellectual output and therefore the digital sovereignty of content creators, we enable a deeper, more profound corruption: the erosion of the shared reality we depend on.

My deepest concern is, what happens when they have all of that data we make publicly accessible? We already know many of these companies are finding new ways of mining data from users of their platforms. Just last week, I read that Meta said it would track employees’ screens, mouse movements, and keystrokes and use the data to train AI.

But what happens when that data isn’t enough? Will companies start targeting the data we hold behind lock and key? Will they want the administrative and restricted data that could further refine their models? Will they begin hammering at those systems next? What will we need to do to protect ourselves and the data we are legally required to protect (and in case you were wondering, yes, I did write that paragraph BEFORE the Canvas data breach)?

Next week we’ll start to talk about philosophy and how important it is to understand the why behind the cultural heritage ecosystem. In the meantime…

• Part 1: Reclaiming Information Sovereignty

• Part 2: The Forcing Function

• Part 3: The Ideological Storm

• Part 4: A Shared Reality

• Part 5: Governance and the Commons

• Part 6: Reclaiming our Seat at the Table

For the last few months, my neighborhood email list has been a hotbed of misinformation (I know, you’re shocked). I have watched people get caught up in narratives that simply didn’t align with the data, history, or reality.

So let me tell you a little story about our school district. My husband is the CIO for our K-12 school district, and we’ve talked over the years about the decline in school enrollment. This is something that I knew about, but being in higher education, I had time before this REALLY started hitting us, my husband less so.

Recently, the news broke that a school MIGHT POTENTIALLY close in my city. The community reacted with the usual drama: yard signs, accusations of mismanagement, and questions about how the school district arrived at the data it supplied. As I watched this unfold from the dinner table, I realized the real tragedy was that people weren’t lifting their heads up from their personalized bubbles. They were fighting over what was changing because they couldn’t see the collective reality.

So what is our collective reality? Let’s start with the “Demographic Cliff.”

In his 2018 book, Demographics and the Demand for Higher Education, Nathan Grawe describes the “Demographic Cliff.” In it, he projects a 15% decline in college-going students between 2025 and 2029 (so like right now).

The convergence of the 2008 Great Recession and the subsequent global pandemic led to a substantial drop in college enrollment that exceeded what Nathan Grawe had foreshadowed in his 2018 research. Despite some recovery in recent years, the growth of traditional bachelor’s degrees is currently being outpaced by the rise in associate degrees and undergraduate certificates (see: Final Fall Enrollment Trends January 2026).

We are quite literally running out of students. My initial reaction to the neighborhood drama was pure indignation. I kept thinking: “How could they be so oblivious? Don’t they follow the news? NPR has an entire series dedicated to the population decline!”

And then I decided to join a class called “Decatur 101” because I thought, “I’m going to be the one to solve misinformation in my city.” (This is the point where you laugh at me because how ridiculous am I?). The class is designed to provide participants with information about the government’s inner workings. In each class session, we engaged with various city departments, listening to presentations about their operations and the challenges they faced.

During Q&A sessions—which often allowed for individual dialogue with city employees—I made a point of asking the same question: “What is the single greatest challenge your department is currently facing?” Most responses were predictably narrow: Parks and Rec lamented a lack of public understanding regarding their specific scope, while Planning cited confusion over local ordinances. Yet, as I processed these localized grievances, a pattern emerged: nearly every department was being blindsided by misinformation without even recognizing it by name.

Did this post spark an idea for you?

Leave a comment

It became clear that the government was failing to communicate the collective context effectively. This was the same breakdown I saw on my neighborhood listserv: a fundamental disconnect in which people couldn’t see that the city’s struggles were often symptomatic of regional shifts or global pressures far beyond the four square miles in which we live.

What I realized was that my neighbors and the city weren’t just being difficult (I mean, some were; don’t get me wrong). They were drowning in a world they lacked the tools to navigate. Ultimately, we are all living in a world with a fundamental breakdown in communication that is preventing us from collecting, processing, and sharing the very knowledge our collective reality requires to maintain society and democracy as we know them.

And then, one Monday morning, I walked into a surprise meeting that was put on my calendar the Friday before. In that meeting, the University Librarian announced her retirement. And then two weeks later, one of my most knowledgeable staff members gave notice of her retirement. Next month? Our library and museum’s Chief Business Officer will retire. And two days before him? One of my go-to counterparts in our Office of Information Technology. By the time these retirements are said and done, collectively these four individuals will walk out the door with over a century of institutional knowledge, leaving us to navigate the gaps they leave behind.

The most daunting realization is that there is no “new generation” waiting in the wings to occupy all of the vacant chairs.

This data from the Economic Policy Institute underscores the reality that our U.S.-born labor force is actively shrinking. Recent projections from the Indeed Hiring Lab show that the U.S. labor force will shrink by 5.9 million workers by 2032. And the sectors getting hit the hardest? Education, healthcare, and government, where retirement rates are already projected to exceed 3%, far outpacing the influx of new talent. When viewed in its entirety, the data paints a daunting picture of the decade ahead.

First, the pipeline of incoming talent is drying up. The steady decline in student enrollment means fewer professionals are being trained to replace those we lose. Second, we are facing an unprecedented exodus of deep institutional knowledge as the number of retirements increases. Finally, the sectors that are deeply human, sectors like education, government, and healthcare, are exactly the ones where humans are most in demand. We face a mismatch: work is growing while the workforce is shrinking.

When viewed through this lens, AI ceases to be a threat to our professional existence and becomes, instead, a lifeboat. But here is the thing about a lifeboat: it prevents you from drowning, but it cannot construct a shore for you to land on. To survive this demographic storm, we must become the architects of the shore that we want to land on. This is the ultimate forcing function.

If you have ideas, please reach out. I’d love to hear what you are thinking or what you’d like me to write about next.

And if you don’t know what a forcing function is, that’s ok because I love a definition. The term actually has an interesting history. It was born in mathematics, specifically in the study of differential equations. There, a forcing function is a function that depends only on time, not on any other variables (thank you, Wikipedia). The term further evolved from Poka-yoke, a Japanese term developed by Shigeo Shingo at Toyota in the 1960s. It vaguely means “idiot proofing”. Decades later, in the late 1980s, cognitive scientist Don Norman used the term forcing function in his book The Design of Everyday Things. There, he used it to mean a strict physical or behavioral constraint that prevents you from continuing down the wrong path. So, as an example, I’m about to microwave a potato for dinner, and I can’t turn on the microwave without closing the door. That is a forcing function. (Again, thank you, Wikipedia, for a great example. Also, my baked potato was delicious.)

The demographic cliff and the silver tsunami are our forcing function. Time is disrupting our equilibrium. We cannot stop the wave of retirements, and we cannot instantly materialize a new generation of workers. At the same time, we have a strict constraint that prevents us from continuing to rely on an endless stream of new hires to sustain our institutions.

The “old way” has been rendered impossible. We have to design the shore of our new world around this forcing function. I know many believe AI is just another platform that technology companies are innovating for the sake of progress (or more accurately, the sake of money), but we are being compelled to adopt these new technologies and rethink our operations. We, the people being impacted by this forcing function, must innovate as an act of survival.

I know I didn’t leave you on a joyful and uplifting note, again. But hopefully, you’ll stick with me; I promise we’re going somewhere good and positive. Next week we’ll (quickly) revisit some topics I’ve already talked about at length — the philosophy of big tech companies, how AI works, and what bot scraping looks like — before getting into new developments in this area that I think we should be keeping an eye on.

Most importantly, thanks for reading The Digital Shift! Subscribe for free to receive new posts and support my work.

Thanks for reading! This post is part of my EBSCO User Group Keynote series, where I’m exploring what it means to reclaim information sovereignty by defining open through the lens of AI. You can explore the rest of the series here:

• Part 1: Reclaiming Information Sovereignty

• Part 2: The Forcing Function

• Part 3: The Ideological Storm

• Part 4: A Shared Reality

• Part 5: Governance and the Commons

• Part 6: Reclaiming our Seat at the Table

I want to start this series with three confessions.

For much of my professional life, I have resided in what I call “The Medium Place.” For those of you familiar with the show The Good Place, you will recognize that the house on the slide below belongs to Mindy St. Claire—the eternal resident of “The Medium Place”. And if you haven’t watched The Good Place, you’re missing out on some great philosophy, but I digress.

I am a librarian; I possess the degree, and I hold the values close (values that often feel undefined, though I hope we can map them out together through this series). But I am also a technologist. And for the better part of my career, these two core identities have been locked in a state of perpetual conflict.

I have spent years in rooms where policy was drafted by librarians who fundamentally misunderstood the technical mechanics, yet refused to learn the language or concepts to bridge that gap. Conversely, I have sat in rooms where technologists have architected policy while completely disregarding the necessity of intellectual sovereignty or the sanctity of the scholarly record.

My career has been shaped by this persistent, nagging desire to belong at tables that never quite seemed to have a chair for someone like me. I realize that might sound contradictory, given my current role at a prominent Research 1 institution, or my history at Stanford, Duke, and George Washington University. Not to mention, I am writing this as if I am an expert. But while my achievements are real, they do not erase the underlying struggle of being a permanent “other.”

If you have ideas, please reach out. I’d love to hear what you are thinking or what you’d like me to write about next.

I have watched library colleagues treat technologists as a support service—as people meant to do their bidding rather than serve as intellectual partners. I’ve lost count of how many times I’ve had to explain that my team possesses expertise far beyond just troubleshooting a laptop. Or how many times I’ve asked folks to learn the new ILS because it is a tool just like Word.

On the flip side, I have watched technologists reject me because I’m a librarian. I have had people from our IT unit try to explain to me how classification schemas work (which is rich, given that our people were creating classification schemas when we were still writing on papyrus, but I digress). I have watched them argue for application security without thinking about the data within the application. Trust me, I could read them the riot act too (and as some of my partners in Emory’s Office of Information Technology know, I frequently do). But today I’m here with you, in a room filled with librarians. So I’m going to focus on what you, as librarians, can do to take back control in a world of AI.

And, worst of all, last week I was told that this dynamic—the tension between being a technologist and a librarian—doesn’t exist, even though my colleagues and I navigate it every single day.

In short, I have been “othered” by both halves of my professional soul, then gaslit into believing the struggle isn’t real.

In her 2016 Ithaka S+R issue brief, Library Leadership for the Digital Age, Deanna Marcum correctly identifies the need for librarians (in this case, librarians in leadership roles) to learn more about digital technologies; however, I think we, as a profession, failed to heed her call. We cannot simply restructure the organization and hope that that makes things better; instead, we must ensure that librarians themselves possess the technical fluency to lead conversations about technology, rather than remain passive recipients of digital strategy.

Librarians who shy away from technical discourse aren’t merely sticking to their expertise; they are retreating into obsolescence and surrendering their influence. We can no longer relegate the need for technical expertise to the periphery of our organizations. Instead, we must place this expertise at the very heart of our strategic planning and organizational growth

Did something I said spark an idea?

Leave a comment

And I have a second confession to make.

I wasn’t always this fired up and certain about AI. In fact, I remember when my friend, the brilliant , first started talking about AI in libraries ages ago; long before many of the people you see on LinkedIn were talking about it (myself included). And I’m going to be honest with you—I didn’t get it. I listened to her because she’s my friend, because she’s smart, and because she’s kind and funny. And, sometimes she’s this little nagging voice in the back of my head telling me to do better. But at the same time, every time I heard her talk about AI, I thought, “Bohyun. AI? Really.”

The slide above shows the first tangible thing I could find she created about AI. It’s a poster she did in February 2018 about AI at Code4Lib in Washington, DC. Based on some quick sleuthing, she had to have submitted this as a proposal sometime in the fall of 2017. For Code4Lib 2018, the community voted on 68 proposals. There were 5 proposals that mentioned machine learning, three of which were accepted to the conference. No proposal — poster or presentation — mentioned AI except for Bohyun’s. Bohyun’s proposal was submitted after she and her University of Rhode Island colleagues had already spent a year developing a library AI Lab. This isn’t to say that there weren’t others talking about or developing AI in libraries, but I do feel its important to point out that in some respects, I and many others are late to the party in relation to folks like Bohyun.

I was also reminded recently that 2018 was around the time Andrew Yang, then an early presidential candidate, was advocating for Universal Basic Income because he thought AI was coming for our jobs.

I want to acknowledge this today. Bohyun, once again, you were right, and I was wrong. doesn’t know me, so he probably doesn’t care about this, but he was also right.

Ok, this is my last confession, and this one is the hardest: I am drowning.

At work, my team currently faces a deficit of four key positions: a manager, a software developer, and two application support analysts. Despite my title as Chief Technology Officer for the libraries and museum, I regularly function as my team’s administrative assistant. I handle everything from processing invoices to booking travel, all while managing a calendar (mine) that oscillates between being overextended and completely unrealistic. Furthermore, I am on call every single night, a necessity born of the fact that I cannot fairly demand constant availability from my only two remaining staff members on call without taking on some of that responsibility myself.

And, of course, because I clearly have a flair for the dramatic, I decided six months ago to start a Substack and accept invitations to speak at conferences all over the world.

WHAT. THE HECK. WAS I THINKING?

I think I’ll conclude here for this week, you know, now that our collective anxiety is through the roof, thinking about all the things I have on my plate.

Next week, we'll discuss strategies for addressing some of the challenges I’ve shared in my confessions, which I know are not unique to me. Together, we'll look at the bigger picture and consider societal changes that make engaging with AI essential.

Last week was a whirlwind of travel for me. After coming home from vacation on Tybee Island, I quickly unpacked my bags and repacked them to head off on another trip. I started the week at the CNI Spring 2026 Membership Meeting and ended it in Toronto for the IFLA symposium, “Being Bold: Clear Direction in a Complex World.”

The transition from CNI’s focus on the how of AI and infrastructure to IFLA’s focus on the why of libraries provided a fascinating yet complementary view of the state of the world. Below are the sessions that truly resonated with me, along with my commentary.

Part I: CNI Spring 2026 Membership Meeting

April 13–14, 2026

For those of you who don’t know, the Coalition for Networked Information (CNI) is a joint initiative of the Association of Research Libraries (ARL) and EDUCAUSE to bridge the gap between libraries and information technology. Established in 1990, it functions as a high-level think tank and collaborative community of over 200 member organizations, ranging from elite research universities to government agencies and technology firms.

CNI has semi-annual member meetings where leaders convene to navigate the complexities of digital scholarship, ensuring that the infrastructure of human knowledge remains accessible, interoperable, and preserved in an increasingly digital world. By hosting these semi-annual meetings and fostering cross-sector partnerships, CNI ensures that the tools and policies used in research today are built on a foundation of collaboration rather than silos.

Opening Plenary: Libraries Leading Campus AI: Claiming Our Seat at the Table

Speakers: Rebekah Cummings (University of Utah), Doralyn Rossmann (Montana State University), et al.

Rebekah Cummings urged the audience to “Lead with library values.” But what are those foundational values? I didn’t walk away from my master’s program with a clear sense of how to define them. My education focused on the technical and mechanical aspects of the profession; I was taught how to describe materials and tweak specific algorithms to build or optimize a search engine.

This gap in formal education regarding values and the philosophy behind them is particularly problematic if libraries want to be positioned as leaders in the ethical deployment of AI across campus. Modern practice demands deep expertise in data privacy, security, and the overarching ethics of technological systems. Without a strong, formal grounding in these concepts from the start, practitioners are often left to bridge the gap while on the job.

Later in the panel discussion, Doralyn Rossmann noted, “We have trust.” But I would question whether or not we do. This isn’t something libraries regularly measure (although I did learn at IFLA that public libraries do). Adding to that are frequent outages caused by AI bot attacks. Not to mention, institutional security has become a critical priority following major cyberattacks on organizations like the British Library. As technology continues to evolve, the profession needs to begin identifying what a measure of trust looks like so we can understand our users’ perceptions and position ourselves for the future.

I have to say, I was really excited to hear from Jessica Davila, the Associate Dean for Digital Strategies and Innovation at the University of Oklahoma Libraries. Jessica and I actually met over Zoom a few months back, and she gave me a fabulous recommendation for one of my favorite podcasts, Close All Tabs (it’s out of KQED in San Francisco, my former NPR station). The work Jessica and the OU Libraries are doing around AI and the policies they’re building is just fantastic. Honestly, I’d encourage everyone to dive deeper into their work. I have been really inspired and have taken some of the things I learned from Jessica to my own institution.

Memory without Origin: The UVA Archival AI Protocol

Speakers: Leo Lo (University of Virginia), Brenda Gunn (University of Virginia)

Frankly, after my Substack post the other week, I had no intention of attending this session. I said my piece, and I was done. But then a number of friends encouraged me to go so there I was in the back of the room. During this session, Leo Lo stated, “This is about disposition, not access.”

I disagree. Establishing minimum provenance standards and contractual language doesn’t mean libraries can prevent AI from taking their content and using it however it pleases. UVA now seems to argue that if you keep things offline, they can’t be taken. Which, sure, don’t put it online, and they can’t steal it, but it misses the bigger picture.

But I think the real discussion regarding archives, libraries, and AI must center on access, authentication, and authorization — who can access the content, are they who they say they are, and what are they authorized to see. Framing this solely as a matter of “disposition” glosses over the immediate technical reality: AI bots are accessing library collections at such high volumes that they destabilize essential infrastructure FOR HUMANS. This surge in non-human traffic often brings systems to their knees, making it functionally impossible for human researchers to access content.

What I once again walked away with is a sense that we must push for better technological literacy within the profession AND we need leadership to articulate exactly why technologists within our profession remain absent from these conversations, even though they are the ones most profoundly impacted by it. And in case you were wondering, yes, my vacation was ruined by what I suspect was a bot attack on one of our vendors.

Sustaining the Public Record: Collaborative Stewardship

Speakers: Merrilee Proffitt (Internet Archive)

The session led by Merrilee Proffitt and the Internet Archive team on the collaborative stewardship of government information was truly exceptional. Their presentation emphasized that preserving government data is vital for maintaining an accessible public record. Inspired by the success of Merrilee’s initiatives, I am eager to find ways for my community to contribute records to Democracy’s Library.

My recent experience taking a “City Government 101” class—where I met with city officials in my own city to better understand the challenges of local government—immediately highlighted the problem of misinformation that Democracy’s Library seeks to alleviate.

Public servants struggle to provide citizens with reliable, contextual information not just on local history and decisions made by local government, but also on the broader societal and demographic shifts that affect municipalities. This lack of access to content and context erodes relational trust within government.

After taking that class and listening to Merrilee speak, I see a significant opportunity for my city government (or state government, for that matter) to engage in broader digitization efforts. By identifying local records suitable for inclusion in Democracy’s Library, we can ensure these municipal records are preserved alongside state and federal documents. This initiative requires more than technical work; it demands proactive engagement with government officials to show that digitization is a necessary step in preparing citizens for informed civic discourse.

Did something I say spark an idea?

Leave a comment

Part II: IFLA “Being Bold” Symposium

April 15, 2026

The International Federation of Library Associations and Institutions (IFLA) is the leading global voice for the library and information profession, representing the interests of library services and their users on the world stage. Founded in 1927 and headquartered in The Hague, it is a non-governmental organization that works in close partnership with UNESCO to promote high standards in library provision, defend intellectual freedom, and ensure equitable access to information.

The purpose of the “Being Bold: Clear Direction in a Complex World” symposium is to address the rapidly evolving global library landscape by recognizing that libraries are being reshaped by complex trends. The event aims to move beyond passive observation toward a state of “futures thinking” as a form of literacy. Through a combination of expert insights and intensive workshopping, the symposium provides a framework for library professionals to identify and test assumptions, plan for uncertain futures, and translate global phenomena into practical, mission-aligned decision-making.

Caught in the Crossfire: Public Libraries, Intellectual Freedom & Culture Wars

Speaker: James L. Turk (Director, Centre for Free Expression)

A staff member recently challenged my post, The Balance of Knowledge, arguing that libraries are never neutral. This really made me stop and reconsider this alongside some of my other writings, and I’ve been finding myself wondering if I’ve actually been contradicting myself! And then I heard James Turk speak at the Symposium.

I now recognize that libraries are not neutral; their curation and resource allocation are inherently driven by values (which are grounded in democratic principles). Yet they remain strictly nonpartisan in providing equitable access to information, ensuring the community can engage with diverse ideas that support intellectual freedom and public discourse.

During his talk, James Turk noted how people often qualify support for free speech with exceptions. This led me to a new realization: I’m one of those people who add a but to the end of my “I’m for free speech” statements. In case you were wondering, my commitment to free speech applies to people, not corporations. I have always questioned whether corporate entities should have Constitutional free speech rights (from the moment the Supreme Court started granting them), a view that aligns with Jimmy Wales’s arguments on organizational nonpartisanship (I’m changing his language here) in The Seven Rules of Trust.

Libraries collect and preserve knowledge as an inherently democratic ideal. Cecilia Preston reminded me at CNI that the ALA was founded in Philadelphia in 1876, the same city where American democracy was born -- and that was intentional. Libraries and democracy are inseparable; therefore, libraries should not be neutral but should actively espouse democratic values. By collecting content on all sides of an issue and providing access without judgment, libraries empower citizens to engage in informed debate.

Exploring the Human–Machine Continent: A Vertiginous Coexistence

Speaker: Ollivier Dyens (McGill University)

Dyens used the language of ‘technology-as-evolution,’ and I began to struggle with his presentation because I questioned whether or not it was a Transhumanist point of view (and if you’ve been reading this substack for a while, you’ll know I hate a transhumanist).

My critique centers on three major points:

1. Innovation Plateau: Dyens argues that technology is evolving. But I think the shift from landlines to cell phones is less revolutionary than the phone’s initial invention. We are experiencing incremental refinement, not true evolution.

2. The Stochastic Parrot: Dyens talks about how profound AI can be. But I also understand the math behind it. And I also have in my head the argument that it is fundamentally a ‘stochastic parrot’ that mimics the user. The apparent ‘coexistence’ of AI and the person Dyens talked about is often a narcissistic feedback loop rather than something novel.

3. Colonial Bias: The ideas AI generates are rooted in Western, non-Indigenous knowledge because it is focused on the content available on the internet, which is, by and large, Western and non-Indigenous. This means that AI becomes a digital extension of colonialism and excludes non-Western epistemologies.

Ultimately, Dyens’ focus on the ‘dizzying nature of the transition’ amounts to an abdication of responsibility when we use AI, preventing critical inquiry into who controls the system.

Since attending his talk, I’ve learned that he falls into the camp of Posthumanists, a branch of philosophy I honestly don’t know enough about to comment on, but I’m definitely looking forward to a deep dive into its roots! Which is to say I reserve the right to change my mind about this session.

Fireside Chat: From Ideas to Practice

Speakers: Pilar Martinez (Edmonton Public Library) and Erin Sniderman (Hampton Falls Free Library)

During the Fireside Chat, the discussion centered on the critical need for upskilling within the library sector. The conversation posed a fundamental question: how can we help the public and staff perceive new technologies as the next evolution of information, much like the transformative shift seen with the advent of the internet? Pilar Martinez emphasized the importance of active participation in this transition, noting that “involvement is commitment.”

A strategic recommendation from conversations Pilar and Erin had with the audience was regarding political engagement. One participant noted that government offices were being visited by some of the more conservative political groups in their community, so the Library chose to focus on being welcoming to these groups when they came by to see what “woke” activity was happening. The result? The library was held up as an example of what libraries should be. By fostering a friendly environment that welcomes diverse viewpoints, even those we might struggle with, libraries can position themselves as essential hubs for civic dialogue rather than being sidelined by political divisions. Jimmy Wales also discusses this concept in The Seven Rules of Trust, noting that organizations risk alienating various groups by adopting partisan stances.

The dialogue also touched upon the library’s historical perception. It was observed that libraries have sometimes struggled to appeal to individuals with more conservative ideologies. This friction often arises because libraries are deeply committed to collecting all kinds of information, including materials that challenge traditional views. Furthermore, the panel noted that the issue of banned books is complex, affecting various political groups and not solely “the left”. In concluding the session, Pilar Martinez reaffirmed the library’s core mission of non-partisanship and intellectual freedom, stating, “We are not promoting the ideas, we are promoting the access to ideas.”

If you have ideas, please reach out. I’d love to hear what you are thinking or what you’d like me to write about next.

Closing Keynote: On Being Brave, Bold, and Right

Speaker: R. David Lankes (University of Texas at Austin)

In his summary of the symposium, R. David Lankes framed the day’s discussions by emphasizing that the future is non-determined and contested, describing it as a “the undiscovered country” (gotta love a Star Trek reference). He stressed that it is ultimately up to us to push towards the future we want to create, urging us to “bend the world toward justice through action.” The future belongs to “a community in the hands of the agile”—or, better yet, to people who need people (well, hello, Barbara Streisand).

Lankes offered a critical view of AI, characterizing it as either a mirror that simply reflects what we have said or a system that “extracts society.” He argued that AI is non-deterministic and lacks a true understanding of the data or content, instead only looking at patterns. This highlights the need to understand data in context and to contextualize data and content by bringing in the local community.

The discussion also touched on significant societal concerns, noting that “we’ve broken down trust.” Lankes challenged attendees to consider how libraries support democracy, what democracy is, and how the creation and curation of knowledge supports it, specifically asking about the values of intellectual freedom. He questioned whether higher education in America has lost trust due to the fact that we have put entire generations into debt, or because it is perceived as “woke”, again echoing concepts around non-partisanship we hear in Jimmy Wales’ The Seven Rules of Trust. Finally, he pushed for librarians to become “radical positive change agents,” asking how we can craft ourselves to be “bigger and bolder.”

Summary Takeaway

The week reminded me that while the technical infrastructure (CNI) is the ground we stand on, it is our willingness to be “Bold” (IFLA) that determines where we go. We must move from being observers of the digital shift to being its architects.

Below is a glimpse of what my unplugged life looks like, including some exclusive photos of my family’s new(ish) puppy!

I highly encourage everyone to find a moment to step away from their keyboard and toward something as serene as my experience here on the coast of Georgia. Also, if you’re looking for a sign to take a vacation, this is it.

Both sides of my family had a profound impact on me. They showed me that a wide network can be a living, breathing infrastructure. Most importantly, they taught me that even when you are upset with one another, moving on and maintaining that network is more valuable than being “right” in isolation.

I’ve been thinking about this family history a lot lately because, as my daughter gets older, I see my small nuclear family taking on the same living infrastructure. Just this morning, my husband and daughter took over some of my usual morning tasks because we were running behind.

My daily support structures mirror a realization I’ve had through writing this Substack: the ideas I share and talk through with all of you are rarely mine and mine alone. They belong to a network of individuals who help shape the ideas I write down, and I cannot, nor would I want to, take credit for all of them. This realization highlights a fundamental tension between how knowledge is actually created—in a web—and how it is often rewarded—as a solo performance.

The Myth of the Lone Expert

In academic, technology, and administrative circles, there is a persistent, ego-driven tendency to amplify one’s own voice while tamping down the voices of others. In higher education, we see this in the tenure and promotion system, which often forces individuals to worry about whose idea is “the best”. In open source, we see this in the elevation of a single developer who gets elevated into the role of benevolent dictator for life.

The idea of a Lone Expert poses a risk to communities. When a leader acts as if their knowledge is singular, they signal that feedback is unwelcome. For a colleague or junior staff member with an alternative viewpoint, approaching them with ideas or critiques is a high-risk maneuver. If the person at the top doesn’t acknowledge feedback or credit others’ ideas, the network stops being a public square for discussion and debate, striving to become the best version of itself, and instead becomes a hierarchy of silence.

Breaking this silence requires a different kind of leadership—one that views admitting ignorance or error as a strength rather than a weakness. Dismantling this hierarchy requires a radical shift in leadership values; it moves us away from performative certainty and toward an environment where admitting ignorance is viewed as an asset. In my own leadership, I’ve found that being willing to say “I don’t know” or “I was wrong” is actually a requirement for a healthy organization. If a leader is infallible, they aren’t a collaborator; they are an obstacle. By modeling vulnerability, we lower the risk of participation for everyone, allowing the network to reward the free exchange of ideas.

Did something I said spark an idea?

Leave a comment

This concept of the Lone Expert also appears in conversations about how we think. In a recent episode of The Ezra Klein Show (gift link), Michael Pollan discussed the difference between “spotlight consciousness”—the narrow, goal-directed focus we use to get through a to-do list—and “lantern consciousness,” the expansive, 360-degree awareness we see in a child. I believe the Lone Expert model is a form of spotlight consciousness; it reduces the flow of ideas to a trickle, keeping a single ego productive.

“The Social is Epistemic”

In Mark Coeckelbergh’s Why AI Undermines Democracy and What To Do About It, he argues that democracy is built on a “social-epistemic basis”—it relies on knowledge, social relations, and trust. Hannah Arendt famously noted that totalitarian systems thrive by isolating individuals from one another and removing them from the public square. Pollan uses this warning to remind us of the dangers of such disconnection.

One of the most eloquent defenses of why networks of ideas matter comes from Lauren Pressley, who recently reflected on social epistemology over on LinkedIn. Drawing on the work of Margaret Egan and Jesse Shera from the 1950s, Pressley argues that knowledge is not produced by individual minds in isolation and then deposited into libraries for safekeeping. Instead, she argues that knowledge is produced through systems of validation, circulation, and critique.

Pressley describes the library not as an “information warehouse”—a place where pre-existing knowledge is stored—but as a social epistemic infrastructure. Using a brilliant metaphor, she suggests:

“The road doesn’t move the car. But without it, the car doesn’t go anywhere useful. The library is the road. But it might be even more accurate to say it’s the whole Department of Transportation, responsible not just for the surface you drive on but for whether the road reaches your neighborhood at all.”

When we treat ideas as the private property of a Lone Expert, we treat knowledge like an item in a warehouse. As Pressley notes, “The social is epistemic.” The network isn’t just a support system for the knowledge; it is a constituent part of the knowledge itself.

When we fixate on the Lone Expert, we effectively erase the living infrastructure that sustains the work. As I’ve navigated my own career, I’ve realized that behind every successful presentation or innovative project is a partner who picked up extra work at home, a manager who (hopefully) cleared the path for the employee, or a team of colleagues who picked up the slack so that their colleague could shine.

In her essay “The Librarians Are Not Okay” explores how passion workers are ground into a “fine pulp of self” by a system that expects them to absorb responsibilities once held by multiple people.

Culture Study

The Librarians Are Not Okay

If you value this work, and this platform, and these perspectives, and find yourself opening this newsletter every week — please consider becoming a paid subscribing member. Your contributions make this work sustainable…

Read more

4 years ago · 156 likes · 26 comments · Anne Helen Petersen

When we fail to credit the network, we are being dishonest about the labor required to do the work. The Lone Expert is often just the person whose network is working the hardest to keep the lights on.

The Democratic Act of Connection

This refusal to acknowledge the network is an attempt to make the labor of the “we” invisible in favor of the status of the “I.” I felt this acutely recently when an administrator told me that I was too technical, a tactic designed to silence expertise that doesn’t fit a predetermined narrative.

This is a classic example of enclosure. By using status or domain expertise to dismiss critiques, authority figures privatize the public square of shared ideas. They build a fence around the network, claiming a monopoly on understanding in order to bypass the messy, collaborative friction that true knowledge requires.

When faced with individuals who try to fence off knowledge and ideas, my response is one of radical generosity. The way we generate ideas and build our professional networks reflects how we value and treat one another. By refusing to work in a silo and instead choosing to elevate my network’s ideas publicly, I am helping transform those private ego-driven enclosures back into a public square where debate and the circulation of ideas improve and enhance everyone’s ideas.

This requires us to consider what Pollan calls our “Consciousness Hygiene.” We, of course, need to be mindful of who and what we allow into our network. You don’t want to follow everyone on social media and assume they’re all experts, but at the same time, you do want to cultivate a network of individuals you trust and who provide the “friction” necessary to make your ideas more resilient.

This week, my own “lantern consciousness” was five specific people: Rachel Frick, Kate Silton, Evviva Weinraub Lajoie, Arran Griffiths, and Tim Shearer. I had conversations with each of them about our ideas, what we’re reading, and frameworks that made my ideas better. Without the five of them, the ideas in this post would not exist in this form. They didn’t just support me; they challenged me and reminded me that the conversations that we have together make a difference. This week, they were my guardrails, ensuring that my ideas and understanding remain part of a vibrant, shared space rather than a solitary echo chamber.

As Lauren Pressley says in her piece, the current moment doesn’t ask libraries to become something new; it asks them to be more fully what Shera and Egan described seven decades ago: a robust epistemic infrastructure. The road matters more now that the vehicles are faster and harder to steer.

If we want to resist the erosion of knowledge, we have to start by being honest about how our ideas come into being. We have to give credit where credit is due—to the thinkers like Lauren Pressley who point us toward intellectual frameworks, to the partners picking up the slack at home, and the friends who tell us “yes, and” when we talk about our challenges and ideas. Trying to be the Lone Expert is a burnout trap. Creating your own intellectual commons is like lashing together a bunch of sticks to create a life raft. When we put them together, we find the camaraderie and pure chaos I found at Christmas Eve dinner with my mom’s family or get-togethers with my dad’s many siblings. And because of these experiences, I will reject the ego of the Lone Expert and choose the life raft that is the network of ideas.

If you have ideas, please reach out. I’d love to hear what you are thinking or what you'd like me to write about next. Don’t believe me? Well, then you might want to scroll back up and read this post again. :)

Last August, I spent a day in Lower Manhattan at a workshop that felt like a glimpse into the big picture of how we produce knowledge today. Hosted by Ithaka S+R and the Apereo Foundation, the Sustaining Open Source Software in the Research Enterprise (SOSSRE) workshop brought together about 40 of us—a mix of open-source lifers, university administrators, research software engineers (RSEs), and funders. For many of us, it was a rare opportunity to engage with one another outside our typical silos, and it served as a powerful reminder that the people who make open source work come from incredibly diverse professional and personal backgrounds.

The goal of the workshop was ambitious: to stop talking about open-source sustainability as an abstract problem and start finding practical, scalable solutions for the software that powers modern research.

As the formal report released today points out, we’ve spent the last twenty years normalizing Open Access publishing and the last ten years making data sharing a standard practice. Yet, the code, standards, and technology that gather, generate, and validate that research are often left out of the conversation.

The Day’s Structure

The day started with an opening plenary that, to my delight, included all women, except for the moderator:

• Patrick Masson, Executive Director, The Apereo Foundation (moderator)

• Cat Allman, VP, Open Source, Digital Science

• Dr. Karmen Condic-Jurkic, Executive Director, Open Molecular Software Foundation

• Clare Dillon, Community Lead, CURIOSS

• Dr. Allison Randal, Senior Researcher, Capabilities Limited

During the conversation, Cat Allman (who, like me, likes to craft with yarn during meetings) offered a metaphor that underscored why we need a nuanced approach to OSS: some software is built to be a hammer, a tool meant to be shared, reused, and passed down. But some research software is legit a Q-tip, purpose-built for a single person or a specific, one-off problem. As Cat noted:

Definitely the best quote from the whole workshop.

We spent the rest of the morning defining the challenges across four dimensions:

• Motivational: Why do people do this work (and how do we keep them doing it)?

• Infrastructural: How do we fund and maintain the “invisible” utilities of the digital age?

• Archival: How do we ensure code is citable and preserved as part of the scholarly record?

• Relational: How do we manage the communities and people behind the code?

In the afternoon, we transitioned from high-level definitions to actionable solutions. The group distilled the morning’s discussions into five specific, high-priority challenges that we then “solved” through a series of lightning rounds and mock grant proposals:

• Defining the Value Proposition: How do we convince universities that prioritizing open-source research software (OSRS) is central to their mission?

• Tracking Usage: How can we accurately identify who is using—or could be using—an OSRS product when the current ecosystem is so frictionless (and therefore invisible)?

• Overcoming Cultural Barriers: How do we shift researcher culture to value “openness” and software maintenance as much as traditional publications?

• Assigning Responsibility: Who should actually be responsible for sustaining this code—the project team, the institution, or the broader research enterprise?

• Identifying Extra-Institutional Support: What kind of external leadership, policy, or funding models are needed to shore up the entire ecosystem?

It was energizing and a reminder of the communal solutions we can identify by diversifying our viewpoints and forcing ourselves to move past “this is hard” to “here is how we fix it.”

My Takeaways vs. The Report

Last August, I shared four key takeaways from the workshop. Reading the formal report today, I was glad to see I wasn’t the only one who found those takeaways valuable. Here is how my original takeaways compare to the report’s deeper analysis.

Takeaway #1: Open Source Software is More Than Code

My Original Thought: We need to stop framing sustainability as just a “developer happiness” problem. It takes an entire ecosystem—quality assurance, product managers, designers, and administrators—to make software thrive.

The Report’s Finding & Recommendation: The report focuses on Relational Sustainability, which emphasizes the management of people and communities. It highlights the routine tasks (maintenance, documentation, and coordination) that often get overlooked because they aren’t enjoyable for volunteers or profitable for researchers. To address this gap, the report directly tackles the afternoon challenges of Overcoming Cultural Barriers and Assigning Responsibility by proposing a systems approach that leverages student labor across various disciplines—not just in CS but also in art, marketing, and business—to meet the non-technical labor needs of software development. Additionally, it recommends that institutions professionalize roles such as Community Managers and Research Software Engineers (RSEs) by establishing stable, secure career paths. It also urges funders to explicitly approve grants for these non-technical roles and calls on scholarly societies to build communities of practice that provide peer support for those managing the social complexities of software leadership. Finally, the report specifically suggests that libraries should take the lead in holding and indexing archives of retired OSRS projects, ensuring they remain part of the citable scholarly record even after the community no longer exists.

Takeaway #2: Open Source is Not a Monolith

My Original Thought: Open source is a patchwork of sub-communities with different norms. What works for a repository system like Fedora might not work for a learning management system like Sakai.

The Report’s Finding & Recommendation: At its heart, Motivational Sustainability focuses on the incentives that drive or stall software production. By distinguishing between foundational “hammer” tools and niche, purpose-built “Q-tip” projects, the report calls for a new standard in OSRS classification. This right-sizing approach helps us normalize the retirement and archiving of software that has served its purpose. Crucially, the report highlights a Value Proposition Paradox: projects often need high adoption to attract investment, but they need investment to reach that level of adoption. To break this Catch-22, the report recommends that funders create distinct funding tracks—offering maintenance-only support for critical but non-innovative foundational tools, while providing separate pathways for experimental research code. This also ensures support aligns with a project’s actual function. The report also introduces “Software Nutrition Labels” or “Badging” to help users evaluate project health at a glance. The idea is that these labels or badges would help address trust and adoption by serving as a form of due diligence for institutional IT departments that are often hesitant to install OSRS because they don’t know how to assess its reliability relative to commercial products. This idea mirrors the classification system Nadia Asparouhova discusses in her book, Working In Public: The making and maintenance of open source software. For those looking for a starting point for creating something like this, check out the badges at Project Type Badges.

Takeaway #3: Mapping the Connections Matters

My Original Thought: Much of our infrastructure is invisible. We rely on standards and shared dependencies that we don’t appreciate until they break.

The Report’s Finding & Recommendation: The report highlights a responsibility gap where administrators often don’t know about this software, let alone which software their researchers rely on. To fix this, the report recommends addressing the challenges of Tracking Usage and Identifying Extra-Institutional Support. One way of doing this might be to appoint a central coordinator to promote open-source and manage institutional policy. However, the report acknowledges that university-level support isn’t always enough. It points to the German Sovereign Tech Agency, which treats open-source software as a matter of national security and public infrastructure, as a successful extra-institutional model for funding the standards and shared utilities that the entire sector depends on. By treating OSRS as public infrastructure, we can move beyond guesswork and conduct stress tests on critical dependencies using tools such as knowledge graphs (e.g., OpenAlex or Dimensions) and IOI’s Infra Finder, ensuring that institutional policy is backed by data. Notably, IOI is already advancing this effort by visualizing overlaps in governance and project dependencies, helping to uncover the hidden networks behind our digital infrastructure.

Takeaway #4: Software is Research (and should be treated like it)

My Original Thought: Libraries have led the way in open access and data sharing, but we’ve dropped the ball on software. We need to treat code as a first-class research output.

The Report’s Finding & Recommendation: This finding centers on Archival Sustainability and serves as the ultimate answer to Overcoming Cultural Barriers and Defining the Value Proposition. The report asserts that software is integral to the scientific method and that reproducibility is compromised without it. To bridge this gap, the report recommends that academic departments update Tenure and Promotion guidelines to formally recognize and reward software development and maintenance as legitimate scholarly contributions. Beyond the institution, the report notes that scholarly societies and publishers bear a significant responsibility for advancing software citation standards alongside article citations. This includes a call for societies to lead joint task forces that redefine what “scholarly contribution” means across disciplines, and for publishers to mandate that code authors receive formal credit in the scholarly record, ensuring software is preserved as a permanent part of it.

Would you love to continue the conversation? Send me a message!

Sustaining the Future

Coming back to these takeaways nearly a year later feels like a return to my own professional origins. While much of my recent writing has focused on the fast-moving (and often opaque) world of AI, this report is a reminder that the “shiny” new tools we obsess over are only as reliable as the open-source foundations on which they are built.

The SOSSRE workshop was a rare moment where the “invisible” people who keep our research infrastructure running were made visible to one another. But as the report makes clear, visibility isn’t enough because the sustainability of critical software is often left to chance, individual heroism, or precarious grant cycles.

The release of this report moves the needle from awareness to accountability. We now have concrete suggestions for next steps that involve every level of the research enterprise.

As I look at my own work, I feel extremely validated. We’re long past due; the research enterprise must formally recognize that software is a foundational and valuable part of scholarship. While mapping our dependencies and professionalizing these roles is significant work, they are a necessary evolution. After seeing the diverse expertise in that room in Lower Manhattan, I’m confident that the communal solutions are within reach, provided higher education, funders, and other organizations adjacent to higher education find the courage to fund and prioritize them.

I encourage you to read the full report here and consider: who at your institution is responsible for the research software you rely on? If you don’t know the answer, that’s exactly where the work begins.

First, science and innovation should push humanity forward without being hijacked by hubris and greed. I’ve personally chosen to opt out of the “cult of personality” that often dominates the tech world, preferring to support companies that value innovation over ego. This personal ethic even informs my daily tools; it’s why I eventually moved from OpenAI to Gemini, seeking a path that feels less like an egocentric power grab. Sam Altman himself once famously said:

To me, Google feels like a corporation selling a product, not a person selling a religion. I don’t want to join a religion; I want to use a tool.

Second, making knowledge accessible to people is far more vital than almost any other human endeavor. We understand who we are by learning where we came from, which is why I prioritize supporting the institutions that facilitate that journey. As Hannah Arendt has said:

This is why I believe so firmly in supporting unbiased journalism—subscribing to newspapers for the hard news they report rather than the opinions they curate. It’s why I donate to PBS and keep it as my default station; just tonight, I found myself captivated by a NOVA episode on Pompeii. Things like this constantly remind me that educating oneself is the ultimate path to freedom. As Arendt reminds us in the quote above, this is precisely why tyranny always seeks to suppress facts first: it knows that an informed public is one that cannot be easily controlled.

The Cobbler’s Rule: Institutional Neutrality

Lately, however, I’ve been thinking about a third core belief: the boundaries of corporate and institutional identity. I am beginning to believe that institutions should step back from taking political or ideological stances unless those stances are directly tied to their core mission. As a die-hard believer in Democratic Socialism, I initially found this a contradiction. How can I advocate for a collective, value-driven society while simultaneously calling for institutional silence? But a (former?) libertarian might have changed my mind.

In his book The Seven Rules of Trust, Wikipedia founder Jimmy Wales argues that to build institutions that last, you must “Create a Clear Purpose” (Rule 3). He talks extensively in the book about how sometimes the facts speak for themselves; they are so damning that it doesn’t require you to take a side.

An interview with Jimmy Wales on the Open To Debate Podcast

To illustrate this, here’s a simple analogy: if a chef is anti-crepe, that’s a professional choice; if a cobbler takes that same stance, it’s just noise that chips away at the trust customers place in his ability to fix shoes. When your purpose is to create food, being anti-crepe makes sense, but a cobbler doesn’t need to weigh in on that debate. This does not, in my view, prevent individuals from championing positive initiatives within these organizations. Rather, it prevents institutions from adopting person-like characteristics by claiming a right to free speech.

This brings us to a difficult tension in the library world. Librarians love neutrality, yet we often feel a moral pull to protect our collections from the ever-hungry machines of those who seek to exploit them. When we allow ideological trends to steer our mandate, we violate Rule 6: “Be Independent.” We risk moving from being facilitators of discovery into moral arbiters of how knowledge should be consumed. While individuals may advocate for change, the institution must remain a reliable, neutral service provider.

For me, a democratic socialist, this neutrality isn’t a lack of conviction; it’s a commitment to ensuring that public infrastructure remains a stable resource for everyone, rather than a megaphone for the group currently holding power. More importantly, it prevents institutions from making ideological justifications for why another community or group must do something (e.g., “Deposit with us! We know how to take care of the stuff.” with the underlying text being that you don’t). Ideological justifications are the same type of justifications colonialists made, which ultimately led to the erasure of indigenous knowledge and the establishment of institutional racism.

The False Promise of Control

This tension between neutrality and protection is exactly why I found myself professionally at odds with the recently posted UVA Archival AI Protocol. While the protocol frames itself as a practical standard, it reads more like a manifesto for digital isolationism. Its core rule—no access without control—relies on a “Principle of Reversibility” borrowed from physical conservation.

In libraries and museums, you don’t perform a repair on an object that can’t be undone. But applying that logic to AI model weights doesn’t make sense. As recently argued for the Authors Alliance, knowledge institutions are being sold a “false promise of control.” However, while Hansen’s critique focuses on the legal fictions created by these frameworks, my concern as a technologist is with the technical ones. I spent months looking for ways to safeguard our collections from large AI companies, but ultimately, I’ve come to realize that once you see something on the internet, it becomes really hard to unsee it.

Remember that an AI model is a mathematical engine that predicts the next most likely token in a sequence. Once a collection is used to adjust those millions of parameters, that information is baked into the model’s understanding. You cannot simply reach in and pluck a specific fact out of the weights without retraining the entire system. Furthermore, most companies rely on WARCs (Web ARChive files) to train models, which, in digital preservation parlance, are considered immutable. While you can technically rewrite a WARC to create a new object, it is labor-intensive, which is why (I would imagine) that neither AI companies nor aggregators like Common Crawl are doing it.

By demanding a “Right to Stop” or a “Right to Decommission” that is technically impossible to enforce, frameworks like UVA’s force stewards to become gatekeepers of a barricade that has already been breached. When we promise a level of control that doesn’t exist—much like the EU’s “right to be forgotten”—we do the world a disservice.

The Beginning of the Governance Pipeline: Creator Consent

This brings us back to the concept of governance. Wikipedia defines it as:

I’ve argued before that openness without governance leads to exploitation. But I like to think of governance not just as one thing, but a pipeline through which decisions move.

The beginning of the pipeline is the conversation we have with creators. Some might argue that granting creators too much control creates an administrative red tape that kills “Open Access.” I disagree. I believe stewardship requires a higher duty of care. As Taylor Swift once noted in an interview:

In other words, the music is a mirror reflecting the listener’s life; once it is out in the world, it changes for better or worse. Creators deserve to know exactly what that means for their work—how it might be morphed or monetized—so they can make an informed choice for how to make it available.

Our role as stewards is to provide total transparency, even if that honesty means they choose not to entrust us with their work. If we can’t offer guarantees that their style or persona won’t be mimicked by a generative agent, we have an ethical obligation to say so. This is a trust-building exercise, not an administrative burden. If the pipeline for materials dries up because we refuse to lie about the “reversibility” of the internet, then that is the price of our integrity.

The Other End of the Governance Pipeline: Infrastructure Costs

On the other end of the pipeline, it is our responsibility to ensure that what we hold remains accessible. This is where my objections to billion-dollar AI companies come in. They are no longer just looking for content; they are using the library as their permanent storage. The reality is that harvesting bots are rarely polite. They do not merely index our collections and depart; they return in relentless cycles, scouring our servers to capture every minor update. I have spent many late nights alongside my team battling ByteDance crawlers that hit our catalog so aggressively they triggered a full-scale application outage. This behavior is, in practice, a denial-of-service attack against the human users we serve. We are being forced to subsidize the computing costs and staff hours required just to keep our digital doors open. As these bot attacks multiply and our expenses mount, we face a future where we may no longer have the capacity to ingest new content. Meanwhile, these multi-billion-dollar entities continue to treat libraries as their own free, permanent storage facility, taking everything while contributing nothing to the infrastructure they exhaust.1

The internet is still trying to figure out which of the two paths it wants to take: the walled garden of CAPTCHA and logins, or a web specifically for bots like the one Wikipedia has built. I still want to build a parallel infrastructure for bots, but let’s be honest: it costs money and time, and there is no guarantee bots will use it. I’ve been thinking that the Oxford Common File Layout (OCFL) might actually be the ideal parallel infrastructure for this—a structured, versioned, and machine-parsable way to present our collections as data without bringing our human-facing apps to their knees. Maybe that’s a post for a future date…

In the meantime, until the companies harvesting our data are willing to fund that infrastructure, we must act to protect it. This is why I am currently evaluating solutions like Cloudflare for my institution (if you don’t know, Cloudflare has a pay-per-crawl beta running). It isn’t about setting up toll gates or ideologically blocking companies; this solution would impose a digital maintenance fee. Just as libraries have historically charged for photocopies to cover labor and materials, we must seek reasonable remuneration for the infrastructure that supports massive AI crawling.

Reclaiming Stewardship: A Path Forward

Ultimately, stewardship in the AI era is not about building walls to hide from technology; it is about building the necessary infrastructure to survive it. If we allow ourselves to be seduced by the false promise of control offered by frameworks like the UVA one, we risk becoming irrelevant gatekeepers of empty buildings. If we allow our systems to be overwhelmed by AI bots without compensation, we risk a total collapse of the digital public square.

I’ve said before that “open by default” isn’t the way we should go anymore. We must move beyond the binary of open or closed and embrace a model of “openness with intention”. This means being honest with our creators about how the digital record can be used, and being firm with commercial entities about the cost of maintaining those collections they wish to harvest. It means recognizing that our core purpose—facilitating the human journey of education—is what deserves our protection, not an ideological attachment to neutrality that erodes our very mission.

As always, I’m curious to hear your thoughts. How do we reconcile the ‘Cobbler’s Rule’ of institutional neutrality with the necessity of a digital maintenance fee for multi-billion-dollar AI entities? Is seeking reasonable remuneration for using our infrastructure an act of necessary stewardship, or does it risk a fundamental betrayal of our commitment to open access? I look forward to hearing your thoughts in the comments below.

Leave a comment

Thanks for reading! This is the last post in my “The Great Rupture in Trust” series. If you missed the earlier parts, you can explore them here:

• Part 1: The End of the Rules-Based Order

• Part 2: The Digital Mirage

• Part 3: The Active Work of Repair (this post)

Over the course of this series, we have explored the collapse of the systems we once naively relied upon. In Part 1, we looked at the macro crisis—the end of the international rules-based order and the rise of transactional, strongman geopolitics. In Part 2, we examined the “Digital Mirage” — how the tech industry’s AI bubble is eroding our shared reality and harvesting information without our consent.

If you are feeling a bit overwhelmed by all of this (and well, the world), just a reminder that it is the natural response to a systemic breakdown. We are living through a profound rupture in the infrastructure of trust. But I don’t want to leave you in a pit of despair.

In psychology, there is a concept known as “Rupture and Repair.“ In any relationship, ruptures (moments of disconnection, betrayal, or conflict) are inevitable. A relationship’s survival doesn’t depend on preventing ruptures, but rather on its capacity to repair them. A bond that survives a rupture and undergoes the active, intentional work of repair actually emerges stronger and more resilient than a relationship that doesn’t.

Which is to say, we can no longer afford naive trust in major world powers (i.e., the global hegemons or huge tech companies in Silicon Valley). The rupture has happened and will continue to expand. But I don’t think we should wait to begin repairing. Instead, we must do the active work of repairing ourselves and our relationships with each other. Doing this work requires effort and a decision to rebuild trust from the bottom up.

In the comments section of last week’s post, Nate Angell brought up a brilliant point. He noted that his and Anna’s goal of renaming AI hallucinations to “mirages” wasn’t just to swap out a word, but to encourage practices for developing our own “intellectual sovereignty.” It’s about a healthy questioning of the frameworks we were handed by these governments and companies.

To achieve that intellectual sovereignty, we must band together—economically, technologically, and socially—to create new, resilient pockets of trust. Here are the three levers I believe we can pull to make that happen.

The Consumer Lever: Voting with Dollars

When we look at the macro-geopolitical landscape, it is easy to feel entirely powerless. The decisions that disrupt global supply chains or erode democratic norms are made in rooms we don’t have access to, and for many, the traditional voting booth feels increasingly disconnected from immediate accountability.

Because of this powerlessness, we are seeing a surge in the public using the one lever they still control: their wallets.

This tactic isn’t new. The Montgomery bus boycotts of the 1950s were a consumer boycott. Because Black citizens made up over 70% of the city’s bus riders, the 381-day refusal to ride severely affected the Montgomery public transit system’s revenue, costing the company between 30,000 and 40,000 fares each day. It was one of the most well-known moments when citizens wielded their collective economic power to starve an unjust system of its profits, demanding change and enforcing accountability when traditional political avenues were closed off to them.

Today, recent large-scale consumer boycotts targeting massive brands like Target and Starbucks are modern symptoms of this same underlying dynamic. As highlighted in a recent PBS NewsHour report, regardless of the specific political catalyst behind any given boycott, the underlying mechanism is the same. Voting with dollars is a crude but incredibly immediate way for the public to fight back against the “weaponized interdependence” we discussed in Part 1.

By strategically withdrawing capital, ordinary people are reminding corporations that their operating license still depends fundamentally on the trust and goodwill of the public.

The Technological Lever: Rescuing Our Content

If the consumer lever is about accountability, the technological lever is about preservation.

In Part 2, we talked about the “AI tax”—the infrastructure costs website owners are forced to pay so that large technology corporations can harvest content for free. In an environment where facts are constantly under threat—whether from AI “mirages” or state-sponsored suppression—we can no longer rely on centralized authorities to act as the infallible stewards of our history.

For a long time, libraries and universities have assumed that federal infrastructure was a safe bet, asking faculty to deposit their research data and publications in subject-specific repositories funded by the U.S. government or societies largely supported by different funders or grants. But as the current administration and the necessity of the Data Rescue Project have taught us, we can no longer blindly rely on the government to maintain these critical tools.

We have already seen alarming examples of governments hiding basic facts or deleting reliable historical resources. Recently, the CIA removed the World Factbook for dozens of countries, effectively erasing a globally relied-upon baseline of data. In an increasingly autocratic global state that does not value reality, centralized state archives become vulnerabilities rather than resources.

Building trust today means recognizing that we cannot outsource preservation to centralized entities. Instead, we need to turn to smaller, more locally based communities that are invested in the content we seek to preserve. By changing our practices, we are actively working to repair trust with communities, ensuring that the next generation has a verified foundation of facts to build upon. Shifting our trust from distant federal institutions to local, invested communities is essential, which brings us perfectly to our final lever.

The Community Lever: The Coalition of the Middle

In Part 1, we discussed Mark Carney’s Davos speech, where he argued that “middle powers” like Canada must band together to survive the clash of global hegemons. To repair the fabric of our society, we must apply this exact concept to the local, individual level.

Autocracy and monopolies thrive on isolation. As historian Ruth Ben-Ghiat points out, authoritarianism succeeds by destroying horizontal trust—the trust citizens have in one another. That destruction leaves individuals feeling alienated and desperate, forcing them to rely on vertical trust in a strongman or a centralized power.

But the threat we face isn’t just a traditional political strongman. In a recent interview on The Opinions podcast, Senator Bernie Sanders was asked why his current messaging focuses on “Fighting Oligarchy” rather than just fighting authoritarianism. His response was telling: “Obviously, I work day and night trying to defeat this guy’s efforts to move us to an authoritarian society. I think the two go hand in glove, by the way.”

He is exactly right. We are rapidly moving toward what Danielle Allen, writing in The Atlantic, describes as a form of “techno-feudalism.” We are yielding our sovereignty to a technology oligarchy that views democratic self-governance as an obsolete nuisance. This new breed of oligarchs prefers a world run like a for-profit joint-stock corporation, where citizens are treated merely as consumers, subscribers, or data points to be harvested.

When the forces of oligarchy and authoritarianism seek to destroy, community building becomes our ultimate act of societal relational repair. But we shouldn’t be aiming to build huge, unwieldy networks. We must prioritize smaller communities. It is within these localized, human-scale spaces that genuine, horizontal trust can actually take root and be sustained, allowing us to exist outside of this techno-feudal system.

Prioritizing small communities doesn’t mean retreating into isolated silos. A true “Coalition of the Middle” requires building robust relationships between these groups. We must start seeing ourselves as representatives of our communities, actively reaching out to other communities to figure out how we can share resources and advance our common goals. By prioritizing marginalized communities, forming local cooperatives, investing in small [fill-in-the-blank], and supporting open or community-governed technology, we actively heal the horizontal, person-to-person trust that the “Great Rupture” broke.

As I wrote in The Cost of Open by Default, we must stop being passive consumers waiting to be saved by the next election or the next technology, and start acting like engaged citizens.

Pockets of High-Trust

We cannot wait for the “rules-based order” to return, nor can we expect Big Tech to self-regulate. We have to build pockets of high-trust—both digital and physical—from the bottom up.

Just as a relationship that survives a rupture and undergoes repair is stronger than one never tested, these new pockets of infrastructure won’t just be a return to the old normal. They will be a vastly superior, more resilient version of society because trust will be something we actively architect, verify, and maintain within our own communities.

Thanks for reading! This concludes “The Great Rupture in Trust” series. If you missed the earlier parts, you can explore them here:

• Part 1: The End of the Rules-Based Order

• Part 2: The Digital Mirage

• Part 3: The Active Work of Repair (this post)

References

Below are a list of items that I utilized while writing the entire series.

Allen, Danielle. “The Problem With $TRUMP.” Ideas. The Atlantic, January 29, 2025. https://www.theatlantic.com/ideas/archive/2025/01/trump-meme-coin/681452/.

Angell, Nate. “Are We Tripping? The Mirage of AI Hallucinations.” Ideas. Nate Angell, February 13, 2025. https://xolotl.org/are-we-tripping/.

Appelbaum, Binyamin. “Economists Clash on Theory, but Will Still Share the Nobel.” Business. The New York Times, October 14, 2013. https://www.nytimes.com/2013/10/15/business/3-american-professors-awarded-nobel-in-economic-sciences.html.

Applebaum, Anne. “Defund Science, Distort Culture, Mock Education.” The Atlantic, January 23, 2026. https://www.theatlantic.com/podcasts/2026/01/reality-reshaped/685289/.

Applebaum, Anne, host. Reality Reshaped. Autocracy in America. January 23, 2026. https://www.theatlantic.com/podcasts/2026/01/reality-reshaped/685289/.

Basilion, Sarah. “The Montgomery Bus Boycott.” Pieces of History, November 23, 2024. https://web.archive.org/web/20241123150734/https://prologue.blogs.archives.gov/2015/11/30/the-montgomery-bus-boycott/.

Carney, Mark. “‘Principled and Pragmatic: Canada’s Path’ Prime Minister Carney Addresses the World Economic Forum Annual Meeting.” In World Economic Forum. January 20, 2026. https://www.pm.gc.ca/en/news/speeches/2026/01/20/principled-and-pragmatic-canadas-path-prime-minister-carney-addresses.

“CIA Terminates Its World Factbook | CNN.” Accessed March 5, 2026. https://www.cnn.com/2026/02/05/us/cia-world-factbook-countries-cec.

Data Rescue Project. “About Data Rescue Project.” February 10, 2025. https://www.datarescueproject.org/about-data-rescue-project/.

Greenwood, Robin, Andrei Shleifer, and Yang You. “Bubbles for Fama.” Journal of Financial Economics 131, no. 1 (2019): 20–43. https://doi.org/10.1016/j.jfineco.2018.09.002.

Havel, Vaclav. “The Power of the Powerless.” The Hannah Arendt Center for Politics and Humanities, n.d. Accessed March 5, 2026. https://hac.bard.edu/amor-mundi/the-power-of-the-powerless-vaclav-havel-2011-12-23.

Klein, Ezra, host. The Most Important Foreign Policy Speech in Years. Produced by Jack McCordick. The Ezra Klein Show. January 27, 2026. https://www.nytimes.com/2026/01/27/opinion/ezra-klein-podcast-henry-farrell.html.

Merchant, Brian. “AI Is the Bubble to Burst Them All.” Tags. Wired, October 27, 2025. https://www.wired.com/story/ai-bubble-will-burst/.

Metz, Rosalyn. “The Cost of Open by Default in the AI Era.” Substack newsletter. The Digital Shift, January 30, 2026. https://rosalynmetz.substack.com/p/the-cost-of-open-by-default-in-the?utm_medium=web.

Mills, Anna, and Nate Angell. “Are We Tripping? The Mirage of AI Hallucinations.” SSRN Scholarly Paper No. 5127162. Social Science Research Network, February 6, 2025. https://doi.org/10.2139/ssrn.5127162.

NBC News. “The AI Boom’s Reliance on Circular Deals Is Raising Fears of a Bubble.” October 6, 2025. https://www.nbcnews.com/business/economy/openai-nvidia-amd-deals-risks-rcna234806.

PBS News Hour Classroom. “Why Target Is Being Boycotted, and How the Company Has Responded.” August 25, 2025. https://www.pbs.org/newshour/classroom/daily-news-lessons/2025/08/pastor-leading-target-boycott-on-its-impact-and-the-retailers-response.

Roose, Kevin, Casey Newton, Rachel Cohn, et al., hosts. Is This an A.I. Bubble? + Meta’s Missing Morals + TikTok Shock Slop. August 22, 2025.

“Rupture and Repair | Psychology Today.” Accessed February 28, 2026. https://www.psychologytoday.com/us/blog/emotion-information/202001/rupture-and-repair.

“Rupture and Repair | Psychology Today.” Accessed March 5, 2026. https://www.psychologytoday.com/us/blog/emotion-information/202001/rupture-and-repair.

Sanders, Bernie, David Leonhardt, and Jillian Weinberger. “Opinion | Bernie Sanders: ‘There Ain’t Much of a Democratic Party.’” November 3, 2025. https://www.nytimes.com/2025/11/03/opinion/bernie-sanders-oligarchs-americas-story.html.

Stix, Gary. “The Science of Economic Bubbles and Busts.” Scientific American, July 1, 2009. https://www.scientificamerican.com/article/the-science-of-economic-bubbles/.

The Montgomery Bus Boycott (U.S. National Park Service). n.d. Accessed March 5, 2026. https://www.nps.gov/articles/montgomery-bus-boycott.htm.

Wikipedia. “Ruth Ben-Ghiat.” February 13, 2026. https://en.wikipedia.org/w/index.php?title=Ruth_Ben-Ghiat&oldid=1338178784.

World Economic Forum. Special Address by Mark Carney, Prime Minister of Canada | World Economic Forum Annual Meeting 2026. 2026. 33:11.

.

I am drawing here on the work of Nate Angell and Anna Mills, who argue that we should stop using the term hallucination to describe AI errors. To say AI hallucinates is to anthropomorphize it. Instead, they propose we all use the term “mirage.” Just as a desert mirage is a real physical phenomenon, an AI mirage is a predictable artifact of how these systems generate content.

This shift in language is vital because it changes where we place responsibility. If we call it a hallucination, we are essentially waiting for AI companies to find a cure for the machine. But if we recognize it as a mirage, the burden shifts to us. It requires us to develop the ability to recognize the shimmer of a statistical guess rather than mistaking it for a reliable point of truth. We have to stop being thirsty travelers in the desert who believe everything we see.

Thanks for reading! This post is part of my series, The Great Rupture in Trust, where I’m exploring how the collapse of global and digital infrastructure is forcing us to rethink how we connect, protect our information, and rebuild community. You can explore the rest of the series here:

• Part 1: The End of the Rules-Based Order

• Part 2: The Digital Mirage (this post)

• Part 3: The Active Work of Repair

The Economics of Bubbles

While the mirage describes the nature of the output, the engine driving its rapid adoption is a classic speculative surge—the kind of “bubble” that historically appears whenever a new technology promises to reshape the world.

But fun fact, the very idea of a bubble is a source of debate in economic circles. Nobel laureate Eugene Fama, for instance, has expressed his skepticism toward the term. In an interview with The New York Times following his 2013 Nobel win, he argued that market price swings are generally efficient reflections of changing intrinsic value. From Fama’s perspective, what we call a bubble is just a myth we invent after a crash to explain away a perfectly rational response to new information.

However, recent academic work has begun to challenge this skepticism with empirical data. In their 2018 paper, “Bubbles for Fama,” researchers Robin Greenwood, Andrei Shleifer, and Yang You analyzed a century of market data and found that the probability of a burst increases significantly when high price growth is accompanied by high volatility and a surge in new stock issuance.

Their approach is further refined by Brent Goldfarb and David Kirsch in their book, Bubbles and Crashes, which focuses specifically on technology bubbles. They identify four core elements that create the perfect storm for the bubble to burst:

1. Uncertainty (about the technology’s ultimate value),

2. Pure Plays (companies focused solely on the new tech1),

3. Novice Investors (drawn in by the hype), and

4. Powerful Narratives (promising a world-changing revolution).

When these four elements come together, they sever the connection between a technology’s market price and its actual usefulness.

Is AI the Ultimate Bubble?

As a technologist who has watched several “next big things” (remember Second Life?) sputter into non-existence, I’ve learned that hype cycles are very real. A recent WIRED analysis categorized the current AI surge as the “Platonic ideal” of a tech bubble, citing it as a perfect 8 out of 8 on the Goldfarb and Kirsch risk scale.

Much like the early pioneers of broadcast radio that Goldfarb and Kirsch discuss in their book, we have been told the value of AI is obvious: it will solve climate change and cure cancer! Yet sustainable business models remain elusive for the industry, and these giant companies are burning through billions of dollars, with costs so high that they effectively lose money every time you enter a prompt into a system. And yet, the powerful narratives these companies have created have successfully drowned out any necessary precautions.

But the most insidious danger here is that the products we are being sold are nothing more than massive autocomplete engines.2 AI prioritizes sounding authoritative and confident over being factually correct. In their rush to dominate the market, the companies producing these technologies end up hiding their extractive nature behind the noble-sounding label of “innovation.” Our personal data, creative work, and history are being harvested without permission. They are turning our lives into raw material for a corporate product.

The trust between the technology’s creator and the user is fundamentally broken.

The Necessity of Digital Sovereignty

If we can’t trust the technology any more than we can trust the countries of the world, we find ourselves in a dangerous situation. This is where the concept of digital sovereignty moves from a boring policy term to a necessity for survival.

As I’ve written before, the open mantra that libraries, cultural heritage organizations, and even the internet have long lived by is falling apart. Anyone running a website today is being forced to pay a literal “AI tax.” When thousands of AI bots scrape a website to train their models, they consume that site’s bandwidth and server power. This effectively forces the website owner to pay the infrastructure bills so AI companies can harvest content for free.3

This is why Prime Minister Mark Carney’s Davos speech resonated so deeply with me. When you’re living in a digital mirage, strategic autonomy means that middle powers (or, in my case, individual communities) must secure their data, just like a nation secures its borders. Digital sovereignty is the realization that if you don’t own the infrastructure, you don’t own your history.

The Information Exit

So here we are, stuck in a “Digital Mirage.” We have been sold the promise of superintelligence, but when we get to it, all we find is extractive practices and a fractured shared reality. When we can no longer distinguish between a verifiable fact and a convincing statistical shimmer on the horizon, our shared reality evaporates.

We cannot wait for AI companies to self-regulate, nor can we wait for countries to create legislation. If trust is infrastructure, then how do we build systems that don’t rely on hegemons?

Next week, in the final part of this series, “The Active Work of Repair,” we will look at how we move from being passive passengers in this rupture to becoming the architects of our own high-trust infrastructure. We will explore how to band together—as a “Coalition of the Middle”—to rebuild trust from the bottom up.

In the mean time….

For centuries, society has operated under an implicit social and political contract (well, hello, Hobbes, Locke, and Rousseau; I thought I would never use you again). We naively believed in a “rules-based order.” We assumed that international trade, diplomatic norms, and basic democratic principles were guardrails the world could count on. But what we are witnessing right now is the collapse of trust.

Thanks for reading! This post is part of my series, The Great Rupture in Trust, where I’m exploring how the collapse of global and digital infrastructure is forcing us to rethink how we connect, protect our information, and rebuild community. You can explore the rest of the series here:

• Part 1: The End of the Rules-Based Order (this post)

• Part 2: The Digital Mirage

• Part 3: The Active Work of Repair

Trust, as it turns out, is not just a warm sentiment (as I’ve written about before). It is scaffolding that allows economies to function, people to invest in technological advancement, communities to thrive, and nations to cooperate. And right now, trust is falling apart, and institutions and people are grasping for power and imposing self-isolation.

To understand how to fix it, we first have to take a step back and look at the macro crisis we are in with a clear head, which is anything but easy to do. But, over the course of this series, which I’m calling “The Great Rupture in Trust,” I will look at why trust is failing, how that shows up in our lives, particularly in technology and information ecosystems, and what we can do to repair it.

Geopolitics Without Rules

Let’s start by looking at the immediate geopolitical topics dominating the headlines. Following the recent World Economic Forum meetings in Davos, the international community is still reeling from U.S. leaders’ speeches that once again called for the seizure of sovereign territories, such as Greenland (I apologize to anyone from Greenland on behalf of all U.S. citizens everywhere). For some in the international community, this felt like a fever dream, but for those of us living in (or next to) the U.S., it felt like old hat. Add in the ongoing and failed weaponization of tariffs against allies (again, I apologize to anyone from Canada on behalf of all U.S. citizens everywhere) and you get, well, the current world order.

This behavior perfectly illustrates a concept political scientist Henry Farrell1 recently discussed on The Ezra Klein Show (gift link): “weaponized interdependence.” For years, we naively assumed that global economic integration would deter conflict. We built global supply chains and financial systems, believing that the world’s police (the U.S.) would use this power for good. What we are all finding out now is that those very networks of integration are now (and have always been) used as leverage against those who don’t fall into line.

Farrell points to the global financial system as the ultimate example. Because the U.S. dollar is THE currency of the global economy, almost every international bank must maintain relationships with U.S. institutions to clear transactions. This allows the United States to turn the entire global banking system into a mechanism of power. The example Farrell gives is when the U.S. ratcheted up pressure on Iran, Iranian banks were abruptly cut off from the system, leaving the country unable to get paid for its oil and forcing it to barter for basic goods like grain and zippers.

Which brings us back to the mood at Davos. What we witnessed was geopolitics without rules. In this new world order, sovereign territories are treated as assets to be acquired in hostile corporate takeovers, and economic partnerships are built on leverage. Davos was a blaring signal to the rest of the world that international relations are governed by whoever happens to hold the biggest stick. And if you hadn’t heard, the U.S. has the biggest stick.

Time to Take Down the Sign

This isn’t just a temporary disruption. In his address at Davos, Canadian Prime Minister Mark Carney explicitly named what we are experiencing. He didn’t call it a transition: he called it a rupture.

There is a vital difference between these two things. A transition is something you plan for—it’s an orderly migration from one version of software to another. A rupture is the 2024 CrowdStrike incident. It was messy and destructive, and no one got to opt out.

Carney, a former central banker, is perhaps the last person you’d expect to deliver a speech that sounds like a manifesto. But he argued that the “pleasant fiction” of the rules-based international order is over. He used the Václav Havel metaphor of the greengrocer—the shopkeeper who puts a “Workers of the world, unite!” sign in his window not because he believes it, but to signal compliance and avoid trouble. Carney accused us all of “living within a lie,” performing the rituals of a global order that we privately know to be false.

While this rules-based order provided vital public goods—open sea lanes, a stable financial system, and collective security—the hegemons always exempted themselves when it suited them. But we have reached a point where integration is a source of subordination rather than mutual benefit.

This brings us to the concept of middle powers. Carney argued that intermediate powers like Canada are not powerless. Instead, they must move away from old, comfortable assumptions to build sovereignty that can withstand the pressure of the hegemons.

In many ways, this is a metaphor for all of us. Just as Canada is a “middle power” caught between global hegemons, we as individuals are “middle powers” caught between the predatory practices of Big Tech and the encroachment of autocracy. Our old assumptions about the open web being a utopia are no longer true. This idea of the middle powers banding together to create a third path is the seed of how we might begin to architect our own pathways for trust.

Autocracy and the COVID Hangover

But before we start to build that pathway, we have to ask ourselves how the infrastructure of global trust rotted so quickly. To be fair, it didn’t happen overnight.

The decay of trust was undeniably creeping in long before 2020, but the COVID-19 pandemic acted as a massive accelerant. The sheer necessity of lockdowns and isolation widened cracks that already existed, damaging our connections to neighbors, local communities, and shared institutions. We became physically and psychologically disconnected from one another at an unprecedented speed.

In a recent interview on the podcast Autocracy in America, historian Ruth Ben-Ghiat brilliantly articulated how isolation provides the perfect fertile ground for authoritarianism. Autocracy thrives precisely by dismantling horizontal trust (the trust citizens have in each other). Once you sever the ties that bind people together, you create a vacuum, a process COVID put into overdrive. That vacuum is then filled by vertical trust—an enforced loyalty to a single leader or state apparatus. When we lose our grip on each other, we reach for a strongman to steady us.

The Illusion of Stability

We can no longer afford the pleasant fiction that the old world is coming back. The trust that held the geopolitical order together is gone, leaving us in a landscape defined by who has the power.

But if we can’t trust the systems that move our money and define our borders, we are forced to rely on the systems that move our information—our digital eyes and ears. The problem, as we’ll see, is that those systems are currently built on a foundation of sand, too.

Next week, in Part 2, we will explore the “Digital Mirage”: the impending burst of the AI bubble (this bit is for you, Todd!), the fight for digital sovereignty, and how the collapse of shared reality is forcing us to rethink where we store our lives.

If you have other ideas or topics you’d like me to cover, please comment below.

As a parent, these milestones usually trigger a sentimental look backward. But this week, as I sat in my living room, the only thing I could focus on was the sound of the refrigerator. It was making that little ding that tells you you’ve left the door open too long. Thanks to a recent software update (yes, it gets software updates, which drives my husband mad), we can now lengthen the time between when the door is opened and when the ding goes off (I submitted this as a feature request!).

It’s a small change, but the sound just now triggered a realization: my daughter lives in a world where even the appliances require an understanding of networked technology. She wears a smartwatch; she uses AI tools to create cute manga pictures that she then traces and colors in; she exists in a digital fluidity that is invisible to her because she has never known a world without it.

And yet, looking at her education over the last decade, I see a huge void. In schools, we teach English, math, and science, and she’s lucky enough to be in a district with a STEM class, but we have completely ignored the architecture of the world she actually inhabits. There is nothing that helps her understand the internet, how it was formed and grew into the “world” we see it to be today. Nothing that teaches her how software is “made”. We are teaching her how to be a passenger in a vehicle that is driving faster and faster, but we aren’t teaching her how the engine works or how to check whether the brakes are cut.

Collective Inaction

In my own profession, libraries and cultural heritage, we are guilty of a similar kind of sleepwalking. As I explored in my recent five-part series, Encountering Collapse, we often cling to the myths of neutrality and openness rather than confronting the power dynamics encoded in our technology. This avoidance seems to be recurring in how cultural heritage organizations respond to AI, too. I see constant suggestions that our challenges are merely “collective action problems.” The sentiment is always: If we just work together, we can solve this.

But too often, this “collective action” is just us talking to ourselves. We treat every technological shift as an administrative problem to be solved by better coordination between libraries. We convince ourselves that if we just build the right consortium, we can manage the change. But this internal focus, or professional navel-gazing if you will, distracts us from the reality that individuals and communities are being exploited right now, regardless of how well our committees are organized.

I just finished Karen Hao’s Empire of AI, which uses the meteoric rise of OpenAI as the narrative backbone to trace the physical and human footprint of artificial intelligence. She reveals the resource extraction and labor exploitation that AI companies use in their race to the top. Hao uses examples from around the globe to show how the AI industry replicates the extractive techniques of the colonial empires of the past (or present?). In the process, she reveals the hidden costs and the powerful forces behind the technology.

After reading it, I realized that my faith in “nice” solutions has been completely dismantled. The “Plundered Earth” chapter was particularly eye-opening. In it, Hao details how companies like OpenAI, Microsoft, and Google have extracted labor, water, power, and minerals from economically disadvantaged countries, just as the colonial powers that preceded them did. She describes a pattern of taking advantage of these communities, stripping them of their land and their sovereignty, while simultaneously releasing press releases about how they are improving them. Individually, these facts weren’t new, but seeing them woven together into a single narrative felt truly overwhelming.

That being said, the final chapter offers a glimmer of hope. In it, Hao profiles Peter Lucas Jones and his work in training AI to understand te reo Māori. Reading this brought me right back to his closing keynote at iPRES 2025. There, he argued powerfully that AI can be done right if we center it on indigenous knowledge and community control. I walked away from that keynote feeling like there might be a path forward. Finding him in Hao’s epilogue only validated that feeling: his work is a concrete example of how we can push back.

But we in the cultural heritage space must be careful because that path isn’t the “collective action” we usually talk about. In libraries and museums, we are so focused on building consortia and large-scale “collaborations” that we miss our role as stewards of the collections of individuals and communities. That isn’t to say these organizations are all bad; we are a community too, and we have needs and wants where collective action may work — open source being one example. But our focus should always be on the individuals and communities we serve and support.

Instead of building more collective action programs and partnerships amongst libraries, we need to focus on supporting the collective action of the communities whose collections we steward. We need to educate them on what we do, the challenges we face in stewarding collections, what makes AI possible (data, labor, energy), and our place in that conversation as cultural heritage organizations. Only by building trust with communities can we effectively engage in collective action, because we must understand the individuals and communities whose needs we seek to meet.

Reclaiming Sovereignty

This brings me to the word we don’t say enough in the United States, but really need to: Sovereignty.

In his speech at Davos, Mark Carney spoke about why sovereignty is critical in a fragmented, unpredictable world. He argued that we can no longer rely on the old rules or benevolent global systems to protect us; instead, we must actively build “strategic autonomy” — the capacity to govern our own futures, resources, and infrastructure without being at the mercy of extractive external powers.

While he was speaking to world leaders, he was also speaking to me, and to you. In the digital age, the communities we serve act as micro-nations. Just as a nation must secure its borders and resources to maintain independence, communities must secure their data and intellectual output to maintain theirs.

For cultural heritage organizations, this means ensuring the communities we partner with retain the ability to understand and make decisions about their own history, data, and knowledge. When we collect their stories but fail to educate them about how the digital landscape works — how technology, and AI in particular, works, and how companies harvest, annotate, and exploit this data to train algorithms and AI models —we are actively participating in stripping them of their sovereignty. We cannot simply archive their past; we must ask them what they need to build their capacity and strategic autonomy. If we don’t, we are enabling those who seek to take away their sovereignty.

The ding of the fridge is no longer just a noise in the kitchen; it is a signal. It’s a reminder that the world has changed, and we have been too slow to keep up.

It is easy to feel defeated by the scale of what Hao describes in Empire of AI. The extraction of resources and the erosion of privacy feel like unstoppable forces. But looking at my daughter, I realize that we don’t have the luxury of defeat. Pessimism might be an honest place to start, but it cannot be where we finish.

We have to move beyond the comforting illusion that simply “collaborating” will save us. We need to do the harder work of building strategic autonomy, both for the communities we serve in our libraries and for the children we raise in our homes.

We need to stop treating technology as magic and start teaching it as machinery. We need to show them the engine, explain the fuel, and warn them about the brakes. We need to stop raising them to be passive users of an empire they don’t understand, and start training them to be sovereign rulers of their own digital lives.

As a technologist, my perspective on openness is rooted in the daily realities of managing systems and protecting the cultural heritage artifacts entrusted to libraries and museums. While the open web has brought incredible progress, we are now facing four distinct pressures that make me question where we draw the line:

1. The Intellectual Harvest of Cultural Heritage: AI companies are no longer just looking for content; they are harvesting the intellectual frameworks we’ve spent decades building. As I’ve noted before, AI companies are not only focused on crawling and harvesting authors’ content. They are also interested in financial gain from the curation carried out by cultural heritage organizations over thousands of years, as well as in the frameworks we have built to describe, relate, and explain the content.

2. The Infrastructure Tax: Scraping bots aren’t always polite. Ironically, while I was drafting this post, my team was battling ByteDance servers crawling our catalog. Ultimately, the bots on ByteDance’s servers caused a full application outage. I’ve said before that these outages constitute a denial-of-service attack against our human users. Anyone running cultural heritage infrastructure on the web is essentially being forced to pay an infrastructure tax (computing costs plus staff time) to keep our digital doors open.

3. The Harvesting of Physical Collections: Recently, the Google Books Library Project has been reinvigorated, and rest assured, it has been reaching out to large research libraries to identify books that they haven’t scanned yet. My understanding is that some libraries have successfully negotiated one-time payments from Google, but in my opinion, this approach is short-sighted. Libraries either forget or don’t understand that these companies don’t just “read” the data once. They revisit it every time they build a new model, effectively getting a perpetual license to our libraries for a one-time entry fee.

4. The Erosion of Trust: Our entire model of stewardship relies on trust. When a donor, writer, or artist hands over their life’s work, they expect libraries and museums to serve as protectors of their works. Currently, we cannot provide them with any true guarantees that their materials won’t be ingested, synthesized, and commodified by AI companies. If we can’t offer guarantees, the pipeline for materials may dry up, and the cultures we seek to document and preserve may be poorer for it.

This last point—the fragile relationship between the donor, the library or museum, and the technology—is exactly what I want to dig into today.

Contractual Barriers to Donor Deposits

I recently reviewed a donor contract that really put point 4 above into focus for me.

Below is a summarized and reworded version of the language I reviewed, based on information from the Author’s Guild, the Society of Authors, and the Australian Society of Authors websites.

Organizations must safeguard materials against their use for:

• Generative AI Development: Training, fine-tuning, or otherwise developing generative artificial intelligence models.

• Mimicry of Authors: Creating outputs such as chatbots, conversational agents, or other simulated responses that emulate, mimic, or synthesize the author’s literary voice, style, or persona.

• Derivative Works: Using the materials to enable the generation of new expressive works that are derived from or substantially similar to the materials.

I should also say that the contract includes positive language allowing us to use AI for internal purposes, such as OCR and metadata creation. But as a technologist, my first read had me concerned that the language creates a significant implementation challenge.

How do technologists adhere to these terms in a world where the internet is essentially an open buffet of content?

The End of Public Availability?

To ensure absolute compliance with the terms outlined above, I see only two things my team can implement to safeguard the materials. And guess what, neither of them feels open:

1. Removing the content from the public internet entirely. I’ve had extensive conversations with my team about our current application that would serve up the material in question, and they are confident that our technical blocks prevent the vast majority of automated crawlers from downloading our materials. I’ve even verified that our content is successfully hidden from the Wayback Machine (which, fun fact, uses data from Common Crawl, which I’ve mentioned before in relation to AI). However, “automated” is the keyword there. If an AI company is motivated enough, it can simply pay humans to build targeted scrapers (there’s a subReddit for that) or maybe vibe code something to bypass these blocks. If the content is on the open web, a determined actor with enough resources will eventually get it.

2. Requiring a legal waiver from every single user. This type of action is a direct response to the shifting legal ground around “meaningful consent.” For years, a case called hiQ Labs, Inc. v. LinkedIn Corporation, 17-16783 (via Courtlistener.com), served as the north star for scrapers; it essentially held that content on the public web could be harvested without legal repercussions. But that legacy has shifted since the AI boom began; courts have started to emphasize that institutions must take “sufficient” and “affirmative steps” to signal that bots are unwelcome. A simple robots.txt file (historically used as a “please don’t” sign) can be easily ignored and is not a legally binding barrier. Without an explicit, user-facing legal waiver that forces a visitor to agree to terms, AI companies can continue to hide behind the ambiguity of public access to claim they have a website’s consent.

This feels like a step backward for cultural heritage institutions. We spent decades trying to tear down the walled gardens around content and criticizing publishers for locking away knowledge. Now, the predatory nature of AI scraping is forcing us to build our own walls.

But maybe there is a middle path? I’ve previously written about emerging standards such as Really Simple Licensing (RSL) and Web Bot Auth. These protocols aim to move beyond the binary of open or blocked by creating enforceable, machine-actionable methods to prohibit bots (or make them pay) from using content on the open web. If one of these standards prevails, we might eventually be able to offer openness with intention, allowing authenticated, responsible crawlers while enforcing the strict prohibitions required by our donors.

Until then, we are essentially waiting for the final word from the landmark lawsuit The New York Times Company v. Microsoft Corporation, 1:23-cv-11195 (via Courtlistener.com). The Times is suing OpenAI/Microsoft, claiming that millions of its articles were used without permission to train AI models, essentially “free-riding” (a common term in the open source universe) on its massive investment in journalism to create a product that now directly competes for its audience’s attention without driving views (and therefore preventing revenue generation through ad dollars).

As of early 2026, the discovery phase has already proven that “regurgitation” (where an AI outputs copyrighted content word-for-word) is a documented reality of generative AI chatbots. This challenges the industry’s defense that AI training is a “transformative” fair use that doesn’t harm the original market. Until the legal dust settles, though, the only way to adhere to the terms I described above may be to keep it under lock and key.

Technical Sufficiency vs. Legal Reality

When I meet with General Counsel next month, I’ll ask: What does “sufficient protection” mean for the university? How much risk is the institution willing to take with our current infrastructure and legal agreements adhering to these terms? What must my team implement to meet the legal thresholds still being established? If a donor is highly litigious, will our “best effort” be enough? Or are we entering an era where high-value digital collections must be siloed behind click-through agreements or worse, stored away never to see the digital light of day?

As always, I’m curious to hear your thoughts. Are we being too protective, or are we the only ones taking the stewardship part of our job seriously?

My connection with Katherine Skinner spans at least 15 years, though for much of that time, I admired her from a distance. I had heard of her accomplishments long before I had the opportunity to cross paths with her. It wasn’t until my former boss at Stanford, Tom Cramer, suggested we connect that the bridge was finally built. When I first started my current position, I found myself literally sifting through the history of her tenure at Emory while cleaning out old files, piecing together how she got her start.

As my own interest in open infrastructure has grown, I’ve had increasing opportunities to interact with Katherine, both personally and professionally. I sincerely appreciate her contributions to our community and, most recently, her work at Invest in Open Infrastructure—so much so that while preparing my recent keynote for iPRES, I made a point of citing her extensive work on community building. As part of that, I decided to reach out and interview her to learn more from her. I also thought it would be a great complement to the talk I was creating (and knew I would post for all of you). Throughout the interview, you’ll hear why Katherine has influenced my work. Which is to say, if you haven’t had the chance to engage with Katherine, you are in for a real treat.

In this conversation, we peel back the layers of “sustainability”—a buzzword we all use but rarely define with the rigor Katherine brings. We discuss the fragility of institutional support, the dangerous “anti-profit” bias in libraries, and why she believes the scarcity of funding in our field is actually an illusion. Katherine’s provides an unvarnished wisdom that I admire greatly. She is a truth-teller in a field that often prefers politeness, and as you’ll read, she’s done apologizing for it.

The Fragility of Institutions

We began by discussing the origins of Educopia and the MetaArchive Cooperative. Katherine’s experience in building these communities serves as a masterclass in risk management: she realized early on that even the most supportive institutions are subject to the whims of administrative turnover. That realization—that “permanent” is a myth—became the bedrock of her philosophy on infrastructure.

RM: I’m very curious about your writing on funding and sustainability. You’ve called it an “illusion” that there isn’t enough money to support open infrastructure. What lived experience led you to that conclusion?

KS: Sustainability, and fiscal sustainability specifically, was always on my mind. When I entered this field around 2000, my first projects focused on archives, OAI-PMH, and building federated collections. But the question that dogged me was: Then what? How do we make sure that these digital records persist?

By 2006, Martin Halbert—who was running library systems at Emory—came to me and said, “We need to found a non-profit.” At the time, we needed a neutral space for a digital preservation network we’d founded (MetaArchive) as well as other community activities hosted and facilitated by our team at Emory, but that we didn’t want to “own” or “control”; we believed these initiatives didn’t really belong to any one institution. We felt that associating with a single institution would be too much of a failure point.

At first, I tried to stand in the way. We were both working on our dissertations at the time and managing about $6 million in grants. I told Martin to go away until we turned in our dissertations. He said, “No, I don’t think we can wait.” So, 19 years ago, Martin talked me into it, and we founded Educopia.

Here’s the thing—and I think I can safely talk about this now—we had perfect support at Emory. I couldn’t have asked for any library to be a better home for multi-institutional collaborative work. But Martin’s point was that libraries are the custodians of long-term memory; we aren’t equipped to manage the business infrastructure needed to maintain it. Emory shouldn’t be in charge of the future of a digital preservation network. That is never going to be Emory’s aim, despite all the supporters we had in the university’s administration at the time. From the get-go, Educopia was a sustainability plan for many of the projects we were working on.

But within a year of founding Educopia, the whole system at Emory started to change, from the Provost down. We lost all of that support apparatus at once. The folks who came in were not nearly as enamored with what we were doing. In the end, Emory was the first one to drop out of MetaArchive.

So, not only did we create a map for fiscal sustainability and governance continuity, but it really, really came true that we needed Educopia for business infrastructure. It left a deep impression on me. You can’t take sustainability for granted. The fragility of the open infrastructure systems we build is often hinged on academic institutions. Some argue that because universities have lasted for centuries, they are a safe bet. But right now? Not a great bet. Not when fiscal models are being upended.

RM: That really speaks to the illusion of stability we often have. You mentioned earlier that there is enough money, but it’s not being used correctly. Can you expand on that?

KS: I started my “Emperor’s New Clothes” series privately in a notebook back in 2016. It was just all the things that I had heard through my project work that frustrated me.

While there has always been enough money in the system to create the open infrastructures we need, there hasn’t been enough to simply replicate the same functions everywhere. We don’t have the money for personality-driven projects, or for one little clique that wants to make it “their way,” so they fork off and do their own thing. The amount of money we’ve had in the system has sometimes let us do foolish things. We’ve had the freedom to do that—money from grants, as well as the volunteer time people had the bandwidth to provide.

Watching that evaporate... well, you gotta find a silver lining. It makes us realize that we’ve been in this age of experimentation, and that it’s been great. We’ve created standards and protocols. But now, we should make some choices. And let’s do that as a field. Let’s not have Funder X decide this program needs to go, while Funder Y pushes another one. We’ve got to find a better way to choose what we genuinely want to work with in the long term. Otherwise, we won’t have any open options because they can’t compete with proprietary solutions. And then we all lose.

Market Failures

Katherine is a self-described truth-teller, and nowhere was this more apparent than in our discussion of the Digital Preservation Network (DPN). As she recounted her experience consulting for the project, her voice offered me a retrospective clarity. DPN wasn’t a story about failed technology; it was a story about ignoring market realities in favor of complex, theoretical constructs.

RM: It reminds me of one of the first open projects I was part of—DPN. I feel like I’ve been in so many meetings and on so many grants where we keep reusing the same technology, but the sustainability piece never gets addressed.

KS: Oh, DPN. I lived through that catastrophic reinvention. Steve Morales hired me as a consultant from Educopia back then. I’m a truth-teller, so I told him, “Your people sent you to me because I’m not gonna mince words.”

Steve was convinced there was a market for digital preservation. And I kept saying, “There is no market. We are trying to build a market for digital preservation, but what you are building is a Byzantine project the community can’t afford.”

He’d say, “I’m getting $20,000 a year from all these institutions. I’ve got a market. It’s secure.” I worked with him for less than a year because he wouldn’t take any of the advice I was giving. I had to cut my losses because he wasn’t interested in what I was telling him.

But you’re right about the replication; DPN itself was similar to other distributed digital preservation services that already existed because the personalities involved thought they needed to build it anew. We have to stop doing that. We have to map our way to success in a changing world.

Risk management requires recognizing that most open infrastructures—whether housed in an academic institution or a non-profit—are operating below their revenue needs. Universities, grant funds, or government funds subsidize them, but this has long been invisible to decision-makers. People thought, “LOCKSS? Stanford is taking care of that. Internet Archive? Other institutions are paying for that, so it’s permanent.”

But none of this is permanent infrastructure. Somebody has to be managing the books behind the scenes. Somebody has to manage community governance. Or else that infrastructure isn’t open or stable and becomes a liability rather than a building block.

Mapping the Landscape: From 2.5% to Infra Finder

Next, Katherine and I pivoted to her current work with IOI. Katherine explained the evolution of Infra Finder, a tool designed to bring transparency to this opaque ecosystem. There was a palpable sense of relief in her voice when she described helping to create a public resource that empowers decision-makers.

RM: I’m curious about the work you’re doing at IOI now, specifically Infra Finder and the granting programs. How do those initiatives help solve these “emperor’s new clothes” problems you’ve identified?

KS: My own role in helping to create Infra Finder actually dates back to a small Mellon-funded grant where we tried to map the scholarly communication landscape. I brought ten years of consulting experience to that project. I had worked with clients whose stories were so much worse than Educopia’s—they put our struggles in perspective.

I realized that the open infrastructure people thought was orchestrated and building toward a crescendo was actually floundering. Everybody was just barely staying alive. I wanted to examine the business and governance practices of these projects to better understand them. I’m a sociologist; I want to see the other elements, not just the code.

We put together an abhorrently long survey that people only suffered to answer because they were my friends. But they answered it confidentially. So then I was stuck with a whole lot of information I couldn’t talk about—the number of groups that didn’t have a budget or anything resembling budget-to-actuals. The number of groups that gave me a list of funders, and I’d have to remind them, “You also got funding from X,” and they’d say, “Oh, that’s right, we did.” They weren’t keeping written records, and it was genuinely scary.

Infra Finder has been influenced by that. IOI created it to bring more visibility to open infrastructures in our field. We needed a neutral space that could look across all these different infrastructures and projects to help organizations using the infrastructure see what was working and what needed work, not just technologically but also from business and community angles. It is designed to help those trying to invest in or adopt infrastructure understand what they are actually getting into.

We now have 134 infrastructures in Infra Finder. We can start looking across the body of data. What are the trends? What are the funding patterns over time? Which of these still has a founder twenty years in, who is holding onto the reins—and is that good or bad?

We need that information to make informed decisions about how to consolidate this vast range of work we’ve built into something we can afford to manage. And I use the word “consolidate,” knowing it’s a loaded term. I’m talking about consolidation with a little “c”—where the community chooses it, where we mindfully decide which of these things serves the purpose well, and how we put more resources behind those few things rather than spreading them thin.

Industry Models: Lessons from Women’s Music

This section was a fascinating detour into Katherine’s academic roots. When I asked her about fostering community, she drew a brilliant parallel between open infrastructure and the “separatist” economies of the women’s movement in the 20th century. It was an excellent metaphor for how marginalized groups build power and how staying on the margins for too long can be a risk in and of itself.

RM: You mentioned your background in sociology. I recently listened to the What Now Podcast, where Trevor Noah interviewed Bernie Sanders, who talked about sponsoring dances and basketball games in Vermont because you need to foster community before you can have political engagement. What do you wish we could promote more of within our community?

KS: My graduate work was on the music industry—specifically the interplay between social movements and new genres. For me, it was the Indigo Girls and Tracy Chapman in my childhood. They made it to Top 40 radio, and it was like, Oh my god, there are lesbians in the world! I’m gonna survive this!

I got to dig into how the women’s movement fostered community through music stages and festivals like the Michigan Womyn’s Music Festival1. There was a sense of autonomy and a “separatist industry.” It resonates with what we are dealing with in open infrastructure right now. We are trying to build something autonomous.

In the women’s movement, it wasn’t that everyone went into a separatist industry—Ladyslipper Music wasn’t the only distributor people used. But the lesbian-feminist movement laid a strong, separatist foundation through bookstores, record labels, and distribution channels by and for women who weren’t being served by the mainstream industry players. You get these stories of empowerment that are both connected to the industry and separate enough from it that they are controlled by a kind of communal sense. People embraced Ani DiFranco and her Righteous Babes label in part because of all it signified politically and economically. There’s a community around her. Ditto for the Indigo Girls, and for some civil rights activists who wound up making music and impacting the development of both musical movements and actual genres. These genres come into interplay with the industry in these really interesting ways.

I think some of what we’re watching happen within our own profession is similar. What we are involved in—this whole open space—is a social movement. Understanding how these networks and communities function is inextricably linked to knowing how to consolidate infrastructure. You don’t want to break the bonds that have been created, but you also don’t want 16 different networks to try to build the same thing. Sometimes resource scarcity can fuel the construction of common pathways. That’s what (I hope) we can see open infrastructure operationalize instead of continuing to sprawl. It has to be something that happens through extensive knowledge exchange, pushing people toward shared norms, just like those early movements did.

The Non-Commercial Trap

From the sociology of music, Katherine pivoted to the complex economics of our field. She challenged the industry’s binary thinking—open vs. commercial—with the nuance of someone who has seen “good” commercial players and “bad” non-profits. Her argument is clear: an organization’s tax status is not a proxy for its values.

RM: That idea of a separatist industry is so interesting. Can you say more about how that translates to the business side of our industry? We often have this hard line in libraries where “commercial” equals bad.

KS: The notion that “open” is inherently non-commercial is a misconception. You absolutely can be commercial and open, and that can be a genuinely radical approach. We see this with examples like RSpace, a commercial company that embraced open practices and continues to operate successfully. This interplay is essential because, regardless of our preferences, we are operating within a capitalist society. The challenge lies in how we build within these existing structures, embed our values in what we create, and use that as leverage to change the mainstream industry.

A core driver of my career has been to incentivize both commercial and non-commercial players to improve their practices. One way to do this is by making their business models more visible. For instance, if an organization claims to be “open” and has open content, we must ask: what happens if the platform fails? Can anyone retrieve that content and act on it? This is where a focus on open infrastructure becomes necessary.

Another common misconception is that open source isn’t commercial. In the hospital industry, a significant portion of open-source activity is commercial. We need to borrow from other industries and explore different models. An unfortunate, somewhat well-earned vilification has made it difficult for open infrastructures that originated in academia to explore various business and governance models, such as co-ops.

I believe the co-op structure, in which a set of institutions co-invests and profits roll back to shareholders, offers one underutilized, powerful, sustainable alternative. This approach is not about suddenly becoming “profit-minded,” but about being business-minded enough to ensure you have sufficient resources to avoid constantly falling into technical debt, which is crucial in the tech industry.

The “anti-profit” stance prevalent in some parts of our values space is deeply problematic. It’s an unexamined assumption that profit is the problem. In reality, certain types of profit are an issue, but we must remember that a non-profit designation is just a tax status; you can have bad non-profits. This rigid thinking, in which certain “illusions” such as “non-commercial equals good” cannot be questioned, hinders our ability to grow.

Two illusions holding back the open infrastructure space are the unquestioning resistance to consolidation and mergers, and the simplistic view that non-commercial is somehow inherently blessed. We need a thoughtful reimagining of what these terms truly mean, how they function, and what we might be holding ourselves back from achieving because of them.

My Takeaway

My conversation with Katherine Skinner was a powerful reinforcement of a core belief I’ve held for a long time: sustainability is not a destination or a passive state of being; it is an active, often tricky, and complex series of choices. This realization moves us beyond the comfortable fantasy that continued existence is an inherent right simply because our mission is noble.

True sustainability demands constant, critical evaluation and the courage to make hard decisions. We see this in major organizational inflection points—the “leap moment” when an entire community decides to embrace evolution, as we witnessed in the significant effort required for the Fedora migration, or the equally difficult but necessary strategic decision to sunset a beloved but ultimately redundant or resource-draining tool. These moments are defined by intentional action, not inertia.

We must fundamentally change our mindset and move beyond the persistent illusion that funding will magically appear just because our cause is just or our intentions are good. Hope is not a budget line item. As Katherine so eloquently summarized, our path forward requires a shift in strategy: we have to stop reinventing the wheel and instead start building a vehicle we can actually afford to drive and maintain. This means prioritizing shared infrastructure, investing in collaborative, scalable solutions, and accepting that pursuing bespoke, duplicated systems is a luxury we can no longer afford. The future of our work depends on making fiscally responsible, strategic choices today that prioritize long-term resilience over short-term comfort.

Thanks for reading! This post is part of my iPRES keynote series, in which I examine the hidden political and cultural structures of our systems. You can explore the rest of the series here:

• Part 1: Encountering Collapse

• Part 2: The Myth of Neutrality

• Part 3: Community as Backbone

• Part 4: The Limits of Openness

• Part 5: Collapse as Opportunity (this post)

Encountering collapse exposes the values embedded in the ecosystem, whether intentionally, as we saw in the keynote from the Honourable Simon Kofe, Minister for Transport, Energy, Communications and Innovation, Government of Tuvalu or unintentionally, as we are seeing in some open source communities (see Ruby on Rails and Ruby Gems for a dramatic view of collapse) — revealing where we have failed. Yet collapse also creates the space for reconfiguration. As Taveras-Dalmau and Katherine Skinner suggest, these moments are not endpoints but thresholds for something new. Each collapse is a chance to encounter the future differently. We can rebuild ecosystems and ground them in our values, as we saw in Minister Kofe’s talk about a digital nation. And in doing so, we can create renewal.

As an American, this topic is near and dear to my heart. I’m convinced that we are heading toward the collapse of our country. I’m anxiously waiting for us to get to year four of the current administration’s tenure. I’m already thinking—and have been for a year now, actually a year ago today — what my country will look like when all of this is over. How will it be rebuilt to ensure that the values that the majority of Americans hold dear continue to live and thrive?

Taveras-Dalmau et al.’s From Paradigm Blindness to Paradigm Shift?

In “From Paradigm Blindness to Paradigm Shift?” Taveras-Dalmau and her coauthors describe paradigm shifts as cyclical: ideas solidify into unquestioned truths, contradictions accumulate, and the structure can no longer hold, ultimately collapsing. The authors call this “paradigm blindness”—our tendency to stay loyal to the familiar even as it fails us.

But the authors also remind us that collapse isn’t the end of a system— it’s also a clearing point when old frameworks lose their grip and new ones begin to surface. It’s the moment when we finally see the ecosystem clearly for everything that it is. In their view, regeneration occurs when that blindness lifts: a new way of seeing that reorients ecosystems toward interdependence rather than exploitation.

What gives me hope in their analysis and the story we heard yesterday is that regeneration doesn’t emerge despite failure—it appears because of it. Out of that exhaustion to preserve the familiar, new patterns emerge: movements centered on collaboration, local knowledge, and care for place. The authors’ Regenerative Paradigm Map captures this perfectly, showing us how ideas, practices, and values that evolve in the spaces collapse leave behind.

However, the authors also warn that regeneration can fall into the same trap as the ecosystems it seeks to replace. The moment it becomes ideology—when we stop questioning its limits—it loses its regenerative capacity. Renewal depends on staying open, adaptive, and willing to confront the discomfort of unmaking what no longer serves.

Katherine Skinner and The Red Queen’s Race

In Katherine Skinner’s “The Red Queen’s Race,” Skinner suggests what the ecosystem should look like for libraries. Skinner argues that academy-owned and academy-governed infrastructures for scholarly communication are trapped in an exhausting cycle of underfunding, fragmentation, and short-term survival strategies that prevent long-term sustainability. They are “running as fast as they can just to stay in place.” (a reference to Alice in Wonderland for those of you that might not be familiar). She identifies seven interlocking causes for this “Red Queen’s race”:

• Chronic underfunding and burnout;

• Prioritization of innovation over maintenance;

• Internal competition for limited resources;

• Fleeting institutional attention;

• Weak business and leadership training;

• Poor assessment and accountability mechanisms; and

• A lack of shared understanding of the total cost of scholarly communication.

Together, these structural weaknesses create a cycle in which open, community-based initiatives must constantly reinvent themselves to remain visible and relevant, yet rarely achieve the financial or organizational stability needed for durability.

In The Red Queen’s Race, Skinner critiques the way academic culture rewards novelty over maintenance, leading to open infrastructure communities that continually chase new grants, features, or recognition rather than investing in the slow, collaborative labor of stability. This imbalance places immense pressure on the people and organizations responsible for maintenance. The absence of shared assessment, funding coordination, and leadership development further weakens these communities. What ultimately falters is not code or hardware, but trust, cooperation, and the social contracts that allow distributed actors to share responsibility for open infrastructure.

Finally, Skinner’s call for collective action situates resilience squarely within networks of relationships. She argues that the future of scholarly infrastructure depends on aligning communities, pooling resources, and establishing governance systems that reinforce mutual support. Her vision is to build interdependent communities capable of sustaining infrastructure through shared investment and stewardship. In this sense, the article strongly supports the idea that open infrastructure should be more than a single community; it should be an ecosystem of communities.

Conclusion

I know you’re probably expecting a story like the previous posts, but this time I’d like to ask you some questions, hoping that together we can create a new story.

How do we, as a community, encounter collapse?

Higher education, scholarly communication, and cultural heritage are undervalued in present-day society. We’ve been working tirelessly to convince the world that preserving truth, memory, and our collective cultural heritage is important. It’s almost humorous how many times I’ve had conversations with people who had this brilliant idea to preserve something or build some technology, only for me to point out to them that someone in libraries or museums has already done it.

So, what if we stopped? What if we changed the way we encounter one another and our profession? What if we tried something new? What if we started by designing our own ending? If we could tear down the organizations that we have built, what would we want to get rid of? How would we reimagine our profession and our organizations? By following the pattern of emergence and regeneration, could we encounter collapse not from a place of fear, but a place of excitement and anticipation? What would it mean to design endings that invite both encounter and renewal?

References

Skinner, Katherine. “Why Are So Many Scholarly Communication Infrastructure Providers Running a Red Queen’s Race?” Educopia, July 23, 2019. https://educopia.org/blog/why-are-so-many-scholarly-communication-infrastructure-providers-running-a-red-queens-race/.

Taveras-Dalmau, Vanessa, Susanne Becken, and Ross Westoby. “From Paradigm Blindness to Paradigm Shift? An Integrative Review and Critical Analysis of the Regenerative Paradigm.” Ambio, ahead of print, September 12, 2025. https://doi.org/10.1007/s13280-025-02232-7.

You can also watch my speech directly via YouTube:

Thanks for reading! This post is part of my iPRES keynote series, in which I examine the hidden political and cultural structures of our systems. You can explore the rest of the series here:

• Part 1: Encountering Collapse

• Part 2: The Myth of Neutrality

• Part 3: Community as Backbone

• Part 4: The Limits of Openness (this post)

• Part 5: Collapse as Opportunity

Libraries often talk about openness as a core tenet of our belief system. For many organizations, making knowledge publicly available isn’t a question; we just do it. But we have embedded openness into our core belief system, leaving the idea without governance. And unfortunately, this can lead to exploitation. Transparency and access must be balanced with consent and collective accountability.

The AI era has revealed the challenges associated with openness. What was intended to democratize knowledge—such as “open access” and “open source”—now risks benefiting exploitative companies that are far from neutral, as I mentioned earlier. Libraries need to evolve from being “open by default” to being “open with thoughtfulness.” This entails creating environments in which agreement, governance, and collaboration guide participation. Genuine openness supports the commons through shared responsibility, rather than through indiscriminate exposure.

Katherine Skinner, The Emperor’s New Clothes

In “The Emperor’s New Clothes”, Katherine Skinner reminds us that openness, on its own, is not a virtue. Without governance—without the frameworks that create accountability—openness quickly turns into exploitation. She calls out the illusions we’ve built around open infrastructure, around progress, and equity. What we call “open” often masks hidden hierarchies, unpaid labor, and decisions made behind closed doors.

Skinner’s analysis of funding patterns exposes how this imbalance plays out. We celebrate grant announcements and new features, but rarely question how the underlying infrastructure stays afloat—or who is asked to hold it up. As she points out, most funding goes toward visible innovation rather than the invisible work of maintenance and community building. That’s the paradox of our current ecosystem: it rewards novelty while depleting the ecosystem that makes openness possible in the first place. It treats shared infrastructure like a free beer rather than a free puppy. In reality, many open projects operate in silos, competing for scarce funds and attention while exhausting the people who keep them running.

The alternative Skinner sketches are less about control and more about stewardship. She points to models like HathiTrust that build openness through shared governance and pooled investment. These ecosystems don’t confuse access with equity; instead, they recognize that sustainable openness requires structure and agreement. In Skinner’s framing, governance keeps communities from burning out or being exploited. If we want openness to endure, we have to treat it as a relationship to be tended, not a product to be consumed.

For a long time, we treated openness as an unqualified good—a moral stance that needed no further justification. But the last few years have shown me that sharing alone doesn’t guarantee equity or sustainability. For me, the ideals of openness have collided with reality. The systems that promise collaboration and shared progress can, without governance, reproduce the same hierarchies they were meant to dismantle.

Nadia Asparouhova, Roads and Bridges

We’ve long celebrated transparency and access as the cornerstones of innovation, but in “Roads and Bridges: The Unseen Labor Behind Our Digital Infrastructure”, Asparouhova forces us to confront the uncomfortable truth that open infrastructure often disguises exploitation. The open-source ecosystems she describes in Roads and Bridges are built by maintainers who are underpaid (if they are paid at all), overextended, and largely unseen. Meanwhile, the companies and institutions that rely on their work reap the rewards. What we call “freedom” in open infrastructure often hides structural neglect. The same openness that powers collaboration also permits exploitation when recognition and consent aren’t part of the equation.

Asparouhova’s core insight is that digital infrastructure, like physical infrastructure, requires maintenance, governance, and long-term investment to remain functional. Yet, unlike roads or utilities, open-source projects lack coordinated stewardship; they are maintained by small, often unpaid communities or individuals. And to make matters worse, economic incentives, venture capital interests, and cultural norms in software development have created a system in which vast social and commercial value is extracted from the unpaid labor of maintainers. Essentially, “openness” becomes a convenient moral cover for the asymmetries baked into the digital economy. Her framing of maintainers as the “unseen labor” of the internet exposes how open infrastructures reproduce the same inequities we claim to transcend by participating in these communities. It’s not a failure of individuals but of a system that celebrates contribution without ever building structures to acknowledge and value the labor that goes into contribution.

The call that emerges from Roads and Bridges is for a new kind of openness—one that balances labor with acknowledgment. Asparouhova asks us to stop equating open infrastructure with virtue and start designing for durability. We need funding models that honor maintenance, institutions that share responsibility, and communities that consent to the load they carry. In other words, the future of openness depends on re-centering the value of labor. If exploitation is what happens when openness runs without governance, then sustainability — real, long-term sustainability — begins when we practice openness while valuing the labor that helps us create it.

Story Time: The impact of AI on Labor

At Emory, we’ve spent years building digital collections and open-access repositories using various open-source infrastructure. But in early to mid-2024, I started noticing more outages across all our systems. We observed a substantial increase in web traffic. It wasn’t from students or researchers, but from obscure devices, and the traffic volume appeared unrealistic. Over and over, thousands of hits, all targeting our digital collections, open-access repository, and catalog.

At first, I thought it was me. Around that time, I assumed responsibility for managing the team responsible for on-call work after a manager left my team. I thought, well, maybe this is normal, and I’ve been oblivious to what is happening. But then, at a conference, a colleague from another institution had to step out of a session because their catalog was down, and we (administrators tasked with overseeing library infrastructure) all remarked that this was happening to many of us recently. I even had a vendor mention they had to increase their security spending because they were constantly under attack. Eventually, I realized that the situation was neither an attack nor a result of my obliviousness; it was simply AI training traffic. I’ve mentioned all of this before, so I will spare you the details, but if you’re new here, you can read all about it in a previous Substack post I wrote.

But the story doesn’t stop there. AI companies already have access to our content in other ways. In this video, I’m showing OpenAI’s chat interface; you can see it retrieving information and images from my institution’s digital collections. How did it get this access? Much of our content is available in the Common Crawl dataset, a massive web scrape that many AI companies use to train their models. So, of course, they have our content—and they are using it to generate information about our collections right now.

Update

A few months ago, after I began drafting this keynote but before I delivered it, I received an email from Google Books asking whether Emory University would like to join the project. I didn’t have the opportunity to discuss this with our University Librarian until December, but in the interim, Google released Gemini 4.

Upon its release, industry experts immediately touted Gemini 4 as one of the best models on the market, and OpenAI saw a significant drop in revenue as users canceled their ChatGPT subscriptions in favor of Gemini (myself included). Watching this unfold, I was reminded of that email from the Google Books project. I began to realize that Gemini’s performance wasn’t magic; it was high-quality data the company had acquired nearly 21 years earlier via the Google Books project.

But the Google Books project isn’t the only high-quality data the company has at its fingertips. Around the same time I received the Google Books inquiry, librarians across the community reported that representatives from Google Scholar were contacting libraries and vendors to ensure Google could easily crawl and index our scholarly repositories. It appears that Google was having difficulty crawling and indexing content in these repositories due to the aforementioned AI bot attacks. For context, most of these repositories contain theses, dissertations, preprints, and open-access journal articles from major research institutions worldwide.

While I don’t have definitive proof that Google is using data from Google Books or Google Scholar to train Gemin, I suspect Google long ago learned the lesson many other AI companies are now fully grasping: the quality of your data dictates the quality of your output. Or in other words, high-quality training data will create a high-quality AI model.

During a recent presentation to library staff—a talk much like this one—an audience member posed a fantastic and challenging question to me: “Why should we care?”

The answer is simple: these companies profit from, and will continue to benefit from, the labor that cultural heritage and knowledge workers have performed for centuries. While I have heard that some libraries have received one-time monetary reimbursements for some of this data, the data these companies acquire from libraries and museums is not a single-use asset; it is revisited, re-leveraged, and re-exploited whenever a company begins training its next AI model.

In addition to exploiting unseen labor in collecting, curating, describing, and maintaining these collections, their online activity to collect our data increases our infrastructure costs. For anyone familiar with technology, you’ll know that the majority of today’s infrastructure is cloud-based and billed using a per-use model. Increased usage directly translates to a higher bill. The more AI companies crawl our websites, the more our costs rise. These problems are not limited to cultural heritage organizations. Read the Docs and Wikipedia are also facing similar challenges.

At some point, something’s got to give.

Conclusion

What does responsible openness look like in a world focused on exploitation? I have extensively discussed labor and its devaluation. We need to return to our roots and reflect on why we valued openness and on our goals for it. In a world where most internet traffic is bots, does it still make sense to keep our materials openly available? If we aim to be ethical in our sharing—meaning we don’t want to provide people’s content to billion-dollar companies so they can become trillion-dollar corporations—who should we focus on, and what changes need to be made to make that possible?

How do we balance the desire to share information with the responsibility to protect it? At Emory, we have collections that focus on the Black experience in America. However, as a predominantly white institution, I personally question whether or not we are best equipped to understand how that content should be protected.

And even for the communities we are equipped to protect, such as our own, we’re struggling to engage in meaningful conversations around how much of our content we should share openly with billion-dollar companies. I also recognize that I am not in the majority on this issue and that restricting access is a radical concept for our profession. Still, I wish I saw more of these conversations in our ecosystem because I fear that our unwillingness to reevaluate this core tenet may lead to our eventual collapse.

References

Asparouhova, Nadia. Roads and Bridges: The Unseen Labor Behind Our Digital Infrastructure. Ford Foundation, 2016. https://www.fordfoundation.org/work/learning/research-reports/roads-and-bridges-the-unseen-labor-behind-our-digital-infrastructure/.

Holscher, Eric. “AI Crawlers Need to Be More Respectful.” Read the Docs, July 25, 2024. https://about.readthedocs.com/blog/2024/07/ai-crawlers-abuse/.

Mueller, Birgit, Chris Danis, Wikimedia Foundation, and Giuseppe Lavagetto, Wikimedia Foundation. “How Crawlers Impact the Operations of the Wikimedia Projects.” Diff, April 1, 2025. https://diff.wikimedia.org/2025/04/01/how-crawlers-impact-the-operations-of-the-wikimedia-projects/.

Skinner, Katherine. “The Emperor’s New Clothes: Common Myths Hindering Open Infrastructure.” Invest in Open Infrastructure, October 29, 2024. https://investinopen.org/blog/the-emperors-new-clothes-common-myths-hindering-open-infrastructure/.