As the semester winds down, many of us are facing questions about generative AI, whether because we’re reading essays written by it, or news articles written about it. As we look forward to the next bargaining session on May 28, we’re still waiting for the administration’s response to our proposal on generative AI, which would protect us and our students from misuse of AI by the administration.
Here, we share a note by Hari Kunzru, a prize-winning novelist and clinical professor in the Creative Writing Program in Arts & Science, about how the fight for protections against AI misuse is part of the fight for academic freedom.
AI and the attack on universities
In a post in March, Sam Altman, the CEO of Open AI, shared a text generated by one of the company’s experimental models in response to the prompt “Please write a metafictional literary short story about AI and grief.” It ignited a new round of speculation about the use of Large Language Models (LLM’s) in creative work, with certain important literary writers claiming that the AI’s text has literary merit, while others score points by underlining phrases that appear to echo canonical sources – including Nabokov’s Pnin.
While it is interesting to speculate about the aesthetic and philosophical questions raised by the increasing sophistication of LLM’s and other kinds of machine learning systems, in some ways it is a distraction from the role of AI in a concerted attack on the cultural and political power of scholars, artists and intellectuals. The most notable current ‘use cases’ for these systems are as tools to sort through text and images to find grants to cancel and members of university communities to deport.
It is in that context Contract Faculty United – UAW is fighting for a contract provision that would establish baseline protections against the misuse of generative AI, including its use by the university administration to develop course descriptions, syllabi, or other instructional material. PSC, the union of academic workers at CUNY, won a similar provision in their recent contract.
Let’s forget, for a moment, ‘intelligence’ and ‘creativity’ – let alone the thorny metaphysical question of ‘consciousness’. These large language models are powerful tools and it matters who controls them and what they’re used for. Their political economy is, to say the least, problematic. The so-called ‘Foundation Models’ on which most applications are based exist as the semi-secret property of the companies that develop them, and their reasoning is extraordinarily difficult to audit.
Machine learning (of various kinds) has amazing potential, from summarizing huge texts to astronomical mapping to protein folding and piecing together ancient papyri or shredded Stasi documents. So why all the focus on machines that can make art? The pauperization of independent artists and writers is part of the same project as the (apparently irrational) destruction of university research. It’s not driven by an economic logic, but by an ideological one – the desire to eradicate one of the few centers of cultural and political power that the US right doesn’t control. The real threat from LLM’s isn’t about whether a model can make a convincing mashup of metafiction and texts about grief. It’s in the anxiety expressed by an NYU linguist in a recent meeting about whether his book proposal about the syntax of pronouns would lead to his work being flagged for suppression by a keyword searching LLM. These tools are being used to launder accountability for absurd and destructive decisions, and we should be judicious about when and where we accept them.
This picture of a right wing project to destroy the higher education sector may seem conspiratorial, but it is supported by evidence. Christopher Rufo, the leader of the campaign against so-called “Critical Race Theory” (now seamlessly rebranded as “DEI”) is probably the most consequential right wing activist of our time. In a recent interview with Ross Douthat of the New York Times, he talks about putting universities into a “state of existential dread”. This is no idle threat, as we are seeing from the targeting of our colleagues at Columbia. The crucial part (linked below) comes at 56:50 when Rufo talks about a plan to deliberately engineer a recession in the sector, to force universities to abandon the current humanities model and adhere to his vision of a classical curriculum based on ‘universal’ ideas of truth and beauty. This conveniently puts an end to critical thought, which will be replaced by an anodyne celebration of ahistorical cultural ‘greatness’. What is being done to Columbia – and potentially to other universities – is in support of this vapid reactionary agenda. A similar project is outlined by Marc Andreessen in a recent episode of the Lex Fridman podcast. Andreessen, who is one of the key players in Silicon Valley’s turn to the far right, characterizes universities as ‘a government funded cartel’. “There is no way to fix these things without replacing them. And there’s no way to replace them without letting them fail.”
The drive to force untested and unreliable AI into libraries, classrooms and newsrooms is part of a project to undermine the ethos of the university and the authority of scholars. NYU’s unionized faculty are acting in the best interest of the institution by demanding clear guidelines on how and when these tools should be used. We should all understand that we are facing an active threat, and we should be united in defending the role and values of higher education.
This summer, we’ll be sharing scholarly, artistic, and intellectual reflections on our bargaining demands and unionization created by CFU members across the university. Have something you want to share? Respond to this email and let us know. And don’t forget: RSVP to observe the May 28 bargaining session in person or by zoom.