It's about nonfiction. (Or at least what's intended to be nonfiction: the article points out that if left unchecked, an AI may make up sources and quotes to reinforce its conclusions. The New York Times reported a couple of months ago that a lawyer who entrusted his research to an AI learned from the judge and the lawyer for his opponent that none of the cases he'd cited actually existed.)

And the program discussed here is intended for large companies looking for "content", at a seven-figure cost.

But Columbia Journalism Review's daily newsletter linked to this New Yorker article in which a writer gave an AI all the articles he'd ever posted, to see what came out when he asked for an article on a given subject. I found it interesting. His ending is disappointingly silly, imagining "Robot Kyle" running amok if he tries to turn it off. There's a legitimate point there in that the megacompany that owns the overall program can keep running "Kyle" articles long after the real Kyle has moved on or been terminated. (Which I think is basically what the current Hollywood writers' strike is about.) But let's not blame it on a malevolent IR.

Anyway: https://www.newyorker.com/culture/infinite-scroll/my-ai-writ... .

Eric