The legendary Canadian author was in Portugal’s second city primarily to discuss her memoir,
Book of Lives, which was published late last year by Penguin, but the discussion, as is often the case with Atwood, was wide-ranging and, naturally, found its way to the topic of Artificial Intelligence.
Atwood told the audience that she had used an AI model once in her life, Anthropic’s Claude. But it wasn’t to assist with her writing. Instead, Atwood joked that she was trying to find out a spoiler about the British detective series
Father Brown.
“Claude gave me the wrong answer, or it lied. Of course, it didn’t know it was lying because it’s not a human being; it’s a large language model,” Atwood said. “It had skimmed and sampled a lot of television reviews, but they never give away the ending in online criticism, so it was misled by the things it had read about the show.”
Atwood ended her AI anecdote by saying the technology, and the way large language models are developed through scraping through previously published works, simply isn’t reliable enough for humans to depend on.
“Human beings are not robots, but they are opportunists, so if there’s an easy way to cheat and it’s hard to detect, people will do it,” she said. “But the thing about AI is that it’s garbage in, garbage out. Even people who use it for business reasons have to check it because it makes mistakes.”
I’d say she has a good sense of the human condition, so she has some authority to make that judgement.
I don’t think AI is garbage, but what do I know. Ehhh, I’m just a regular jane doe.
Nerdlinger on
I mean, what isn’t garbage in, garbage out?
Other than an incinerator, I guess.
Euphoric-Taro-6231 on
A woman of her time it seems.
blackandbluegirltalk on
If I ask Claude about a spoiler he tells me he can’t access the show and only has the synopsis online. For TV shows with a full *transcript* online, he *can* give me a spoiler if I tell him to search the web. I don’t think she’s lying, but I don’t think it’s the gotcha that she thinks it is. ChatGPT for sure will give ten people seven different answers, and guess what, all of that is getting fed back into the machine because reddit posts are training data.
>
Even people who use it for business reasons have to check it because it makes mistakes.
It’s an imperfect technology that was released to the public to fuck around with and exploit, and it’s still being built! I don’t understand why anyone **would** trust it blindly if it put their job or their life on the line – but that’s really the people she’s talking to.
5 Comments
The legendary Canadian author was in Portugal’s second city primarily to discuss her memoir,
Book of Lives, which was published late last year by Penguin, but the discussion, as is often the case with Atwood, was wide-ranging and, naturally, found its way to the topic of Artificial Intelligence.
Atwood told the audience that she had used an AI model once in her life, Anthropic’s Claude. But it wasn’t to assist with her writing. Instead, Atwood joked that she was trying to find out a spoiler about the British detective series
Father Brown.
“Claude gave me the wrong answer, or it lied. Of course, it didn’t know it was lying because it’s not a human being; it’s a large language model,” Atwood said. “It had skimmed and sampled a lot of television reviews, but they never give away the ending in online criticism, so it was misled by the things it had read about the show.”
Atwood ended her AI anecdote by saying the technology, and the way large language models are developed through scraping through previously published works, simply isn’t reliable enough for humans to depend on.
“Human beings are not robots, but they are opportunists, so if there’s an easy way to cheat and it’s hard to detect, people will do it,” she said. “But the thing about AI is that it’s garbage in, garbage out. Even people who use it for business reasons have to check it because it makes mistakes.”
How many books has she published? And more, apparently: [Full Bibliography – Margaret Atwood](https://margaretatwood.ca/full-bibliography-2/)
I’d say she has a good sense of the human condition, so she has some authority to make that judgement.
I don’t think AI is garbage, but what do I know. Ehhh, I’m just a regular jane doe.
I mean, what isn’t garbage in, garbage out?
Other than an incinerator, I guess.
A woman of her time it seems.
If I ask Claude about a spoiler he tells me he can’t access the show and only has the synopsis online. For TV shows with a full *transcript* online, he *can* give me a spoiler if I tell him to search the web. I don’t think she’s lying, but I don’t think it’s the gotcha that she thinks it is. ChatGPT for sure will give ten people seven different answers, and guess what, all of that is getting fed back into the machine because reddit posts are training data.
>
Even people who use it for business reasons have to check it because it makes mistakes.
It’s an imperfect technology that was released to the public to fuck around with and exploit, and it’s still being built! I don’t understand why anyone **would** trust it blindly if it put their job or their life on the line – but that’s really the people she’s talking to.