Okay, Should Information Really Be THAT Open?
Okay, Should Information Really Be THAT Open?
Localized knowledge, information sovereignty, and the AI menace to open access
It’s time for the collective shock around AI to settle into a clear, critical consciousness. It’s not about whether it’s a good tool that we should all embrace, or a bad tool we should hate and get rid of. Politically, the crisis of data exploitation for the training of AI models should lead us to question the nature of information itself, including the "information commons" and currently prevailing western counter-attitudes that “information wants to be free.” We can keep pushing strict regulations for AI companies, supporting AI product boycotts, and agitating for workers’ rights in affected creative fields, but let’s also pull back for a second and recognize that the problem lies with the basic institution of private property, and that this is an issue with pretty serious colonial dimensions.
My Book's Gonna Get Stolen By Rich Assholes
This week, I’m spending more money than I’ve ever spent on anything in my life (besides like, my college degree or my car [RIP on both counts]) to produce an audiobook. I’m doing this because audiobooks are a great joy and passion of my life, and because a broader commitment to accessibility and literacy kind of demands at least a consideration of alternative media. As an early career librarian with a specific interest in multimedia, too, it feels like an obvious decision.
(Someday I’ll do a full post on the prevalence of un-talked-about reading disabilities in the general public and the importance of non-print formats. Like, does your ability to concentrate also feel completely fucked after three or four covid infections? Mine does. I’m not a doctor, but this sounds like a disability that specifically affects one’s ability to read. Maybe greater access to high-quality alternative text formats would help us all to read more.)
Anyway, the audiobook’s worth it, but dropping stacks on an intellectual property feels very different than dropping stacks on, like, a car. Both the cost and the effort of the audiobook production have me thinking about what all of this is for. Like, where and how people will create value out of this work. I’m talking about use-value mainly, i.e. reading the book and getting something out of it — but potentially also exchange-value, i.e. money.
I plan to post the e-book version of the project at a pay-what-you-want rate, aka free. I will do this because I wrote the book to radicalize liberals against shitty urban design and I really, really want people to read it and get something out of it. Distribution plans aren’t totally clear for the audiobook yet, but I picture doing the same thing eventually for the audio production as well. I’m excited for people to read my stuff for free. This is the beauty of open access!
The ugly part of open access is the possibility that this thing I spent most of my adult life working on will get picked up as part of a digital free library, treated as a tiny morsel in a delicious database meal, and gobbled up to train a large language model somewhere. Literally March 20, the day I was gonna publish this post, Alex Reisner published an article in The Atlantic titled “The Unbelievable Scale of AI’s Pirated-Books Problem.” He cites court documents that show a senior manager at Meta claiming, “books are actually more important than web data” for the training of large language models. Uh oh! Documents obtained by the court also show that Meta employees totally knew what they were doing was wrong, encouraging each other to stay quiet on the quote-unquote “medium-high legal risk” of scraping LibGen for pirated books without compensation to authors.
The court case in question, and resulting discourse, only confirms what I and other authors knew already: Our copyrights don't matter to tech corporations. The operators of these AI models literally do not care at all. In fact, they’re pushing hard for this kind of theft to fall under fair use, and they might win.
In a recent edition of her newsletter [citation needed], Web journalist Molly White offers some thoughtful backlash to the initial defensiveness of independent creators facing the theft of their work by AI scraping. To White, AI mass data exploitation is a serious problem, but one that we should not attempt to solve by attaching locks and limits to our works on the open net. Acknowledging the exploitative nature of AI scraping, she clarifies:
“The solutions [authors] are reaching for — more restrictive licenses, paywalls, or not publishing at all — risk destroying the very commons they originally set out to build. . . Instead of worrying about “wait, not like that”, I think we need to reframe the conversation to “wait, not only like that” or “wait, not in ways that threaten open access itself”.”
As an access tool that allows users to more easily fill their information needs, AI doesn’t in itself constitute a threat to the commons, White argues. Rather, it’s exploitative AI companies that threaten open information by abusing vital infrastructures like Wikimedia, and by flagrantly failing to follow legal guidelines for attribution and use of any scraped media, regardless of they're open access or not. “Even if AI companies don’t care about the benefit to the common good,” she writes, “it shouldn’t be hard for them to understand that by bleeding these projects dry, they are destroying their own food supply.” She argues for solutions that involve “forc[ing] AI companies to engage with these repositories on their creators' terms”, mainly by expanding existing laws to establish fair structures around attribution, consent, and compensation for creative labor.
While I follow Molly White up to a point, I think that — as in all situations where lefties in the imperial core try to develop critical consciousness on an emerging issue — it’s important to go deeper into how we understand the existing legal structures we’re trying to reform. She makes some essential points about the ableism and classism inherent in measures some creators have taken to protect their work from misappropriation by powerful companies. More Captchas and paywalls on the creator-end don’t represent an adequate solution. Yet I think it’s important to empathize more strongly with the individual artists, authors, and Wikimedians who are watching their shit get misappropriated by rich assholes. It’s clear to me that the notion even open information should come with strong, maybe even compromising limits and controls — what White describes as the “wait, not like that” approach to internet publishing — actually has a lot of wisdom. Especially when we consider knowledge creation in a colonial context.
![A screenshot of an article, "'Wait, not like that': Free and open access in the age of generative AI" with the byline "The real threat isn't AI using open knowledge -- it's AI companies killing the projects that make knowledge free." Posted with tag Artificial Intelligence on 14 March 2025 to [citation needed] a newsletter by Molly White. A thumbnail image shows a vampire biting a laptop with Wikipedia open on its display.](https://evieippolito.com/assets/5574945731021034383-unnamed.png)
Open Access to Information Is Neither Neutral Nor Infallible
First we have to historicize the open access movement, to place it in an actual time and place. In 2025 it seems almost a given for every western leftist, or even liberal, to support the general free use and spread of information. Ask a Jeffersonian Democrat and a US anarcho-communist what they think on the subject and they’ll probably both agree that growing and maintaining the information commons is a pretty good idea for everyday people. White opens her argument with the Wikimedia vision statement, positing a world where “every single human being can freely share in the sum of all knowledge.” Whether as a utopian vision or a general guiding principle, you could do a lot worse than believe that, generally, people should be allowed to read and know stuff.
But this is far from neutral, and not inherently liberating. Kimberly Christen, a settler academic I follow who works in the field of Indigenous knowledge organization, argues in her article “Does Information Really Want to be Free? Indigenous Knowledge Systems and the Question of Openness” that,
“The [contemporary] celebration of openness, something that began as a reaction to corporate greed and the legal straightjacketing of creative works, has resulted in a limited vocabulary with which to discuss the ethical and cultural parameters of information circulation and access in the digital realm.”
The valuing of open information and easy access to knowledge, greatly expanded since the invention of the internet, has been a constant theme in US cultural and intellectual history. But like any orthodoxy, this ideology based in a particular set of assumptions that are historically determined, and not beyond reproach. Dichotomies like open versus closed, or public versus private, are rarely based in reality and are, instead, a matter of interpretation and belief. This US approach to information is very different from traditional Indigenous ways of knowing, which while pluralistic and differing, often involve systems of reciprocity, obligation, and relationship in their creation and spread of information.
It follows then that, in Christen’s words, “Commons were never free, nor did they promote an unregulated notion of freedom.” She charts the legal and philosophical origins of public domain in the US back to westward expansion and the displacement of Indigenous peoples. As a legal and epistemological structure, public domain formalized knowledge that was free to use, but it did so in explicit service to US settlers, not the peoples whose homes, lives, and cultures they stripped away. In defining the commons, public domain also enclosed them. It is no wonder that to this day, so many Indigenous cultural artifacts and traditions, as well as artistic works and scientific knowledge, are considered free to use in the public domain, without need for proper attribution or compensation to the colonized people who create them.
It is easy to imagine any number of everyday situations where the access and spread of information is restricted, not for profit or to reproduce or enforce unjust hierarchies, but for entirely good or even liberatory purposes. In my own context, for example, I might pick from any number of community spaces and nights out that literally keep trans women alive, which gradually become defunct as they are discovered and dominated first by white queers, then by the cis and straight general public. There is a reason why “gatekeeping” has become a meme word with positive rather than negative connotations in queer NYC. My favorite queer repository, the Lesbian Herstory Archives in Park Slope BK, holds a collection of unpublished papers collected over the last thirty-five years. Archivists maintain a strong connection with the creators of archived materials, recording attribution guidelines and usage rights based on their individual wishes; a great deal of materials remain private to the archive in this way, undigitized and accessed almost exclusively by queer researchers on-site. And some people in urban unhoused communities, while they might readily share knowledge with other ingroup members, may well not want their names, lifestyles, places, or methods of survival inscribed in maps or writing, ever.
These should not be thought of as minor exceptions to the basically solid baseline of the Wikimedia creed, that “every single human being can freely share in the sum of all knowledge,” for the basic fact that all of the examples above are part of the sum of human knowledge. And they should absolutely not be “freely shared.” When we take general entitlement over all human knowledge as a given — even with certain exceptions and special conditions — we ignore the unspoken political nature of what and whom information is for. Information is expressed, repressed, propagated, and transformed according to inherently political processes, usually governed by a dominant oppressor class with its own ruling systems of government and knowledge production. I don't see any good reason to reduce our treatment of colonized and oppressed knowledge to a choice between either bending and morphing it to fit within dominant western frameworks of law or morality, or disregarding it altogether. This points to a need for a new way to think about that phrase “all human knowledge.”
Returning to the question of AI and open access, it's clear this kind of secret information is not exactly what Molly White was discussing in her article. Nor is it what open access creators and archivists typically have in mind when they create information for the web. However, the line is not as solid as we might hope between open information in the commons, and localized, cultural or in-group-specific knowledge. Personalized, relationally transmitted information, created by and often “gatekept” for the benefit of the communities it serves, very often crosses that permeable barrier into the information commons, at which time it’s called folklore or youth culture. But its free use is not free. Especially once on the internet, it does not — and often cannot — come with fair, material compensation for the people that created it. Black trans women, for example, rarely enjoy legal property rights over the art, cultural trends, and behaviors they create that go on to massively influence all of popular culture when appropriated by white people.
Who could blame creators in such a hostile environment for wanting to put strong and limiting conditions on the use of their particular knowledge, when it so easily passes into the hands of the powerful and leaves them with nothing in return?
There’s no easy way around this. Open access to many kinds of information remains a worthy goal, and also, truly open access to all human knowledge is not the liberated future we might imagine it to be. Neither is it even possible, since some of the most useful and important knowledge exists in the context of land, place, and human relationships, isn’t necessarily static, and cannot always be represented digitally.
You might say, “Well that’s not the kind of information that’s implied in the idea of an open access internet.” And like, yes. That’s exactly the contradiction. If what I’m saying about certain forms of knowledge being invalid for open access sounds obvious, I agree. It begs the question though, why does rhetoric around open access still so often fail to make this distinction? Why do AI enjoyers — both companies and their users — claim that existing models can answer any question we might have? Why is it totally normal to observe that the internet contains “all human knowledge?” That we can learn any skill on YouTube? Pirate all media? Look up any person, place, or thing with a Google search?
Re-Distributing Information Sovereignty
In an information space where knowledge is routinely appropriated from the less powerful for the profit and exploitation of the powerful — now more massively than ever with the proliferation of AI — the dream of absolute open access and free knowledge is not the radical future it’s so often framed to be. For many Indigenous, Black, and queer ways of knowing, it never has been. The answer is not to regress by making more things private, but rather to expand our information politics and literacy beyond the gravitational center of AI, the open net, and digital media altogether. A liberatory politics of information means creating new localized structures of authority and information, focused on the needs and knowledge-ways of marginalized groups, pluralistic and not necessarily open as a given rule. Our theory and our methods have to move beyond the old axiom that “information wants to be free” and start centering instead the freedom of people who create and use information.
Like Molly White, at the current moment I am utterly hostile towards AI companies, and I’m usually on one or the other side of general skepticism towards the general potential of AI as a tool. Its uses in medicine seem cool. Its application as a technology of war, repression, and genocide is disgusting. As an information professional and complete AI laywoman, I’m prepared to learn more about it with an open mind. But none of us can afford to be naive about where it comes from or the regime it currently serves. Like, of course it's a vampire, right? Draining resources from the commons and using up its own food source through mass data misuse? That’s pretty par for the course for the machinery of industrial capitalism. AI in its current iteration should be thought of mainly as an ecological disaster and a weapon of the state. Its limited applications as a consumer tool have mixed but interesting implications for research, and like, insurgent tactics. But I am far from convinced that anything about AI is inherently democratizing, let alone liberating.
What AI does do, is it changes the information landscape in a huge way. We ignore this at our own peril. For independent and marginalized authors, artists, and creators, the decision to publish digital content more carefully, with more barriers to access — or simply to publish less on the open net — is completely sound and sometimes a very good idea. I have far less optimism than tech journalist Molly White regarding the ability of the law to construct and enforce guardrails on what AI companies can do. I would encourage any reader who still harbors great hope in the corrective potential of the state to tame the Ai’s wild growth to consider just how permissive the current US administration has promised to be with corporate regulations, and how much the state stands to gain in military and surveillance capabilities by cozying up to tech specifically. Couple the threat of corporate misappropriation of user IP with the additional threat of an expanding and increasingly hostile surveillance state and, to any group of people historically oppressed by rich assholes, the information commons is looking a lot less inviting right now than well-meaning Wikimedians might hope.
While maintaining and defending the free resources created by the open access movement over the last thirty-or-so years, we have to now consolidate our power where it’s strongest: in our local contexts. To some degree, this means dividing, and potentially balkanizing, the dream of a united free internet. That may not be such a bad thing. The idea that all knowledge comes from, lives in, or is accessible through internet search, AI, or any other digital means, has never been true nor should it ever be.
Let’s take stock of the information resources we have available in our own cities, neighborhoods, congregations, and subcultures. Instead of resigning ourselves to the appropriation of our creative works or compromising our creative needs and desires for the dubious promise of AI, let’s focus on ways to obviate the need for it by expanding library networks, teaching critical information literacy, and nurturing local publishing efforts both formal and informal, to create thriving information economies in every local context. Let’s talk about what it would look like for our local businesses and organizations to divest from corporate cloud computing and storage. Let’s foster technological dual power. Let's expand services like Mesh NYC and invest in localized, communally operated server infrastructure. Let’s make it kinda weird and impolite to say “Hey Google” into your phone.
Let’s adjust how we imagine authoritative sources of knowledge in the first place. Let’s recognize and help build structures of Indigenous information sovereignty. Let’s stop assuming we should have immediate access to everything we want to know and embrace a more intentional, sometimes slower approach to gaining worthwhile knowledge.
Let’s teach art, writing, and communication in ways that emphasize the sorts of knowledge and expression that can’t just be mimicked by ChatGPT, so that middle and high schoolers feel less inclined to use ChatGPT to think for them. While we’re at it, let’s develop critical pedagogies to teach people of all ages how to read full-length books again in great numbers, rather than blaming their inability to do so on readers themselves or public school teachers. Let’s create and propagate local backups of online libraries and other open access resources as they exist now, so people have access to them in the event that fascists succeed in their effort to obliterate them completely.
This is obviously hard, but I’m starting to feel like it should be. Becoming a librarian, I’m learning that equal to the question of information access and availability are the questions of information literacy and information sovereignty. It’s not enough to just ensure that information is there for the taking. Optimistic efforts to create new legal guardrails, compensation methods, and rights of attribution — while important — cannot negate the reality that the information economy is inextricable from racial capitalism, colonialism, and imperialism. Liberation demands going further. Tools themselves don’t democratize anything, revolutions do.
thanks for reading a long one! and thanks to tech queen Arya Burke for her comments on the WIP. i've been thinking about this for a long time. let me list here some blogs and projects that helped me form my thoughts:
Mukurtu CMS, an open-source Indigenous knowledge organization effort originating through a collaborative effort between Waramungu elders and settler designers, used by a bunch of different institutions to host cultural heritage materials online
Mesh NYC <3 <3 <3
"There Is No AI Revolution" by Ed Zitron, on the AI market bubble
"We Need To Rewild The Internet" by Maria Farrell and Robert Berjon
"I Will Fucking Piledrive You If You Mention AI Again" on Nikhil Suresh's blog Ludicity
catch me vending my zines at Purgatory BK March 31st at Protest Fest a diy show for Trans Day of Visibility. write to me at eviewrites@duck.com. subscribe to these posts by newsletter or find me on Substack & IG & Bsky @everzines. with love,
your e-neighbor in digital cyberspace,
evergreen<3