‘Handmaid’s Tale’ Author Margaret Atwood Warns AI Poses Threats Beyond Comprehension

Margaret Atwood, author of the dystopian science-fiction classic The Handmaid's Tale, has warned that artificial intelligence poses a huge and unknowable threat to humanity. 

Speaking at the Budapest Forum, an annual international policy convention, this week, Atwood said the risks posed by AI cannot be accounted for yet "because they’re so new."

"If it were science fiction writers going to town on it, you would get something like 2001, in which the computer decides to kill all the human beings because it doesn't feel they're necessary," Atwood said. "These are the kinds of things that haunt people's nightmares."

Atwood has voiced these concerns before, albeit on a more personal scale. She is among the signatories on an open letter calling for tech companies to compensate writers whose works have been used to train AI models. 

For all their existential threat, Atwood added at the forum that AI chatbots are "terrible" writers compared to human authors.

"They can't even do punctuation. And one of the reasons, of course, is that they've been trained on a huge amount of a verbal output, some of which is quite faulty… and factually incorrect. So you can't even depend on them to tell the truth."

Yet Atwood said the technology’s current limitations actually highlight the need for human supervision. 

"You're still going to need human fact checkers, researchers and copy editors because these things can't [do this function]."