When "To be or not to be?" is a prompt and not the question
Toward increasingly paranoid readings of digital media.
It’s my 41st birthday, and since I officially get one grouchy fortysomething wish, it’s for readers to consider this: Artificial intelligence is an anti-community technology whose principal civic innovation is to reduce social trust and engender a greater sense of unreality in our lives.
A few weeks ago, I read an essay from somebody in the tech business who compared the rhetoric around AI’s current capabilities to Mao’s Great Leap Forward.1 I shared the piece with a couple friends and mentioned how much I enjoy it when I encounter somebody from a non-humanities career track who’s a lively writer. Of course, as soon as I shared the compliment, I realized the essay could have been written in part or in whole by AI itself. As a proud 100% organic writer, I’d have no choice but to Immediately Kill Myself. Just one thing stood in the way of me reclaiming my honor: I had no way of knowing if the writer used AI.
…In seriousness, the real victim of AI here wasn’t my own pride as an English major™ (who cares), but the increasingly fragile ways we engage each other through media.
Not all innovations reduce costs. AI increasingly pushes non-users to engage each other through a prism of doubt, almost evoking Eve Sedgwick’s famous description of the “paranoid reading” style from queer theory: “In a world where no one need be delusional to find evidence of systemic oppression, to theorize out of anything but a paranoid critical stance has come to seem naive, pious or complaisant.”2 Assailed by this unwelcome technology, readers and writers have responded by witch-hunting AI infiltrators using detection software that I would not place my own faith in.3 An open culture, an ethical culture, requires a presumption of innocence even when in the presence of guilt. But I can’t deny AI makes this presumption more challenging.
Using AI for writing is not like traditional plagiarism, where proof of the crime is right there on the page. If somebody writes “To be or not to be?” in the AI era, you won’t know if that was the question or the prompt. I go on LinkedIn and have no idea which super-poster executives might have been rotting their brains on AI all day (though I feel like I have a pretty good idea). The most recently available Princeton University senior survey from 2025 found that “the number who used ChatGPT on an assignment when it was not allowed rose to 27.7 percent, up by 12.5 percent from the Class of 2024.”4 But you don’t know which one out of every four students it was.
Journalism is at least slightly more immunized from AI sludge in that facts, once stated, can be checked: The writer either interviewed the subject (if the subject was real) or they didn’t. But this has not stopped prominent publications from getting tripped up by “journalists” who were AI confections just lifelike enough to be taken for the real thing.
Increasingly the desire for confection is the point. In a coming preprint paper titled “Towards a Post-Social Media Studies,”5 Petter Törnberg and Richard Rogers of University of Amsterdam touched on this broader trend of dehumanizing the human media:
Whereas 52% of the US population reported to talk to AI chatbots in 2025 (Elon University News Bureau, 2025), only 35% of the US population reported to have posted on Facebook, and 20% on Instagram, and 11% on Twitter/X (Gallup, Inc., 2023). Hundreds of millions of users now engage with these systems not merely as tools for completing tasks, but as primary interlocutors: for news and information, for advice, for creative collaboration, and for conversation, including forms of social and emotional engagement previously associated with both media consumption and human relationships.
It’s an earnest embrace of AI’s surreality that is in some ways more challenging to cope with than the lying, because at least a liar knows they’re doing something wrong. Around L.A., I’ve been seeing cartoonishly AI-illustrated billboards for the reality TV star Spencer Pratt’s mayoral campaign, who lost his Pacific Palisades home in the 2025 wildfires and has apparently been spamming social media with AI videos as well. Pratt, currently one of the top candidates, is running to take over an L.A. that exists in his head but not in mine, and this passage from Vanity Fair is like a capsule study for the post-media era to explain how we can live in the same city without our realities ever intercepting:
He showed me his average daily screen time on his iPhone: 16 hours and 49 minutes. (None of that time is spent listening to podcasts or reading news publications. “I don’t listen to anything or read anything,” he says. “I haven’t watched a single thing on my TV since it burned down.”)6
Feelings can be honest, but only facts can be true. It’s against these conditions of unreality from which the writer, the journalist, the artist, the judge, the academic are supposed to stand apart, even when it’s lonely. “The political function of the storyteller — historian or novelist — is to teach acceptance of things as they are,” Hannah Arendt wrote. “Out of this acceptance, which can also be called truthfulness, arises the faculty of judgment — that, again in Isak Dinesen’s words, ‘at the end we shall be privileged to view, and review, it — and that is what is named the day of judgment.’ ” The rest of us might call it Election Day.
Healthy democracies have to support the separateness of fact-finders even while the actual practice of democracy subjects citizens to being “canvassed by a multitude of applicants, who seek to deceive him in a thousand different ways, but who instruct him by their deceit,” as Alexis de Tocqueville put it. In the democratic tradition, AI is another supplicating deceiver for us to learn from, but not necessarily believe.
happy belated!
Happy Birthday Matt!