Information Access as a Public Good
Learning from Librarians, Libraries and Library Science
By Emily
Yesterday I posted about how the silly failures of Google's "AI Overviews" feature throw into high relief the way that chatbots as information access systems rupture the relationship between author and reader. I urged readers to consider returning to the idea that information access systems should be a public good. I think this is a really important idea.
The way I wrote that post, though, bought into a narrative that is false and one I'd rather not be helping to spread, namely that Google's vision of information access has supplanted other systems. I wrote:
Taking the long view, [...] I think it's time to reconsider information access as a public good and to remember that as recently as the 1990s, that's generally how it worked. You wanted information and didn't know where to start? You went to the library.
That sure makes it sound like I think libraries are outdated relics of the past, of limited relevance in the modern world. In fact, libraries (and librarians and library science) continue to be hugely relevant, though (in the US, at least) they are also currently both underfunded and under attack by forces who hope to neutralize their democratizing function.
On top of that, libraries, librarians and library science are the place to look for deeply informed ideas and practices around what it means for information access to be a public good. For example, in "Digitizing Books, Obscuring Women’s Work: Google Books, Librarians, and Ideologies of Access" Anna Lauren Hoffman and Raina Bloom contrast the ideology of "access" behind Google Books—access is a technological problem, with technological solutions that will serve everyone uniformly—with the "community-oriented, care, and service-centered ideology of access" that informs the practice of librarians. (They also provide really interesting discussion about the ways in which this work is gendered, specifically feminized in contrast to the masculinity associated with software engineering work. )
I've long been frustrated at the way that tech bros claim to be "democratizing" things like "access to AI" or "access to information". It always seems to be deployed as an excuse for hoarding data and compute in enormous, toxic piles; the claim that it's okay because now "everyone" can use the result to extrude synthetic text and images. My rejoinder has been: "Democracy means shared governance, not just broad access. Where is your system for allowing everyone represented in that data to have a say about how it is collected, stored, and used?"
But reading Hoffmann and Bloom's piece, I see that there's a whole other salient dimension that I have been missing. Democracy means public, open spaces where civic workers are there to help members of the public navigate to and make sense of information that they need. Democracy means connected communities where people work with one another to understand the world around themselves. A search engine can be a useful tool, but Google's irrelationality goes back at least to their conception of their mission as "organiz[ing] the world's information and mak[ing] it universally accessible and useful." As Hoffmann and Bloom convincingly argue, the very framing of that goal as something that can be achieved by one company, with impersonal technology, is inconsistent with meaningful access for many (most?) of the people in the world.
So, as we look for ways to repair our information ecosystem in the wake of the depredations of the current AI hype-driven deluge of synthetic media, a key step should be shoring up support for library systems and learning from the practices, traditions, and current research of librarians and library and information scientists.