This appears to be the conclusion drawn from a recent spate of reports from the world of journalism. Last month, my colleague Maxwell Zeff wrote about writers who generate at least some of their prose through unlined AI assistants. The star of their article was Alex Heath, a tech reporter who said he regularly has AI writing drafts based on his notes, interview transcripts, and emails. That same week, The Wall Street Journal profiled Fortune reporter Nick Lichtenberg, who told the newspaper that he relies heavily on AI to further his work. Since July he has written 600 stories; One day last February, he had seven bylines.
Since reading these reports – thankfully produced by human hands – I’m having trouble sleeping. Until recently, the general consensus was that it was rhetorical to use large language models to create truly professional prose. Many publications, including WIRED, have strict guidelines against AI-generated text. We also don’t use it for editing, which is less worrisome, though still a troublesome practice for many of the others cited in Zeff’s column. The book publishing world, trying to protect itself from the avalanche of self-published slop, is still monitoring its inventory; Hachette Book Group recently withdrew a novel that apparently relied heavily on LLM’s output. But as models turn into prose and it becomes increasingly difficult to distinguish them from human output, the convenience and cost savings of using AI for the hard work of writing are threatening to go mainstream. The walls have started breaking.
As one might expect, many people were unhappy to read about this development, especially those like me whose keyboard is dripping with blood. But the subjects of the stories are not backing down. It seems as if they believe the future is in their favor. When I contacted Heath – whose work I respect – he confirmed that he had received the push, but shrugged it off. “I see AI as a tool,” he says. “I don’t see it as a replacement for anything – the only thing that’s been replaced is hard work that I didn’t want to do anyway.”
Of course, for people like me, the hard work of writing is an important aspect of the whole endeavor, which lends itself to the task of communicating effectively and clearly. Heath thinks he connects with readers through his writing – he says he’s trained his AI to listen like him, and his Substack includes personally written information about what he’s doing. On the other hand, he told me that since he talked to Zeff, he has almost “one-shotted” some of his columns. “When I say all at once, I mean I didn’t have to do almost anything,” he says. But Heath rejects the idea that letting AI write prose for him means he has bypassed the thinking process, which many believe can only happen through actual writing. “I’m just getting rid of that very dirty, painful, blank page from the void,” he says.
The Fortune writer, who was the subject of the Journal article, has also faced backlash not only from the public but also from his friends and colleagues. “I am feeling the strain in close and personal relationships,” Lichtenberg admitted in an interview with the Reuters Institute for the Study of Journalism. In an email, Alison Shontell, Fortune’s editor-in-chief, tried to dissuade me from the idea that AI is taking the jobs of journalists under their watch. “the important thing is [Lichtenberg] It is not being used as a writing replacement,” he wrote. “Their stories are written with the help of AI versus written by AI. “He’s still doing a lot of ambitious reporting and analysis and rework that is extremely original.”
<a href