Should Writers Be Afraid of AI Writing Better Than Them?
A meditation on the coming world
On a recent episode of Daniel Oppenheimer’s excellent Eminent Americans podcast, an assistant computer scientist at Stony Brook University, Tuhin Chakrabarty, discussed his research on creativity and AI. Chakrabarty’s AI model can produce writing that is very close to a human author’s, and it can be difficult to tell what is what. This is, for many writers, a source of genuine anxiety. “To be clear, Tuhin isn’t claiming that AI, except under the very specific, rather contrived circumstances of his experiments, can already match the best human writers,” Oppenheimer writes. “And he doesn’t know if they ever will be able to match or surpass us. I don’t know how much consolation that affords, though; what he has done is threatening enough.”
I would consider myself an AI skeptic. I do not think the business model of the industry makes a great deal of sense and it worries me that so much of the American economy is now tethered to what might be a financial bubble. The data centers required to power the supposed AI revolution are deeply disruptive to local communities, tax the energy supply, and can pollute waterways. I worry about the effects of AI on children, the loss of critical thinking, and the slop that’s now proliferating online. I am not a doomer about AI in the sense that I am doubtful there’s any great civilizational crisis that will arise from its growth. The graver threats remain climate change, nuclear war, and the conventional warfare we see in the Middle East, Africa, and Ukraine.
But I will concede the AI evangelists at least one point. The technology will improve, and there will come a point where an AI like Claude can write a passable novel—maybe, even, a very good one. Fed enough data, a machine will always get better at mimicry. Perhaps, over the years, it will grow more creative, too. If AI will never fully supplant the journalist, a human talking to other humans, or the columnist, who is paid to offer commentary that fundamentally derives from their own human perspective, creative endeavors can be properly replicated. Poems, novels, paintings, film—the machines, or AI, can do them. They might, one day, graduate from the slop phase. This has, understandably, caused a good deal of anxiousness among artists. There are writers terrified of the day when AI can write as well as them—for some mediocrities, Claude and ChatGPT are already there—and make them redundant. I do believe AI will improve to the point where it can produce, on its own, literature, or something resembling literature. The models will learn, more data will be absorbed, and the sentences will look nice, and the stories will make sense.
Am I, as a writer who shuns AI, feeling anxious about this? The answer, as I’ll explain, is unequivocal.
No.
