As a person who ultimately hates doing tasks, I really want AI to work. I want AI to relieve me of the monotony of emails and Slacks and folding laundry so that I may instead spend my time swimming in lakes and reading novels.
But as a writer who values her work, I am also terrified of AI, for all the reasons laid out by my colleague Clare Duffy's story about the author Jane Friedman.
Here's the situation: Friedman says that one of her fans found several books on Amazon that appeared to be written by Friedman. They had similar titles and subjects. But they read like crappy imitations.
"It was just obvious to me that it had been mostly, if not entirely, AI-generated," she said.
Artificial intelligence tools like ChatGPT are essentially big content vacuums that suck up text from the internet and then, when prompted, regurgitate that information in way that is eerily convincing. In theory, you could ask a bot to write a love story in the style of Allison Morrow's Nightcap newsletter and it'd probably do a decent job imitating me. (Side note: please don't report me to the bots, I really need this gig.)
And sure, often you can squint and see that the AI version has gotten a fact or two wrong, or made a leap in logic. But those bugs are getting worked out, and soon even the telltale AI quirks may impossible to spot.
Amazon removed the fakes posing as Friedman's and said the incident is under investigation. So that's something.
But I, like Friedman, worry about a whack-a-mole approach that relies on writers to report their own impostors.
"What's frightening is that this can happen to anyone with a name that has reputation, status, demand that someone sees a way to profit off of," she said.
Read Clare's full story for more on what authors are doing to protect their work.
Comments
Post a Comment