Margaret Atwood, author of The Handmaid's Tale and The Blind Assassin, says she used Anthropic's Claude exactly once and walked away unconvinced after the chatbot gave her incorrect information about the British detective series Father Brown. Speaking at the Babell Literary and Cultural Festival in Porto, Portugal, Atwood described the experience as the AI producing what amounted to falsehoods — while making clear she understood why: the system has no mechanism for knowing when it is wrong.
What Atwood Said, and What She Was Looking For
Atwood was searching for information about Father Brown, the British detective series, when Claude returned an answer she found inaccurate. According to a recap by Deadline, she put it plainly: Claude "gave me the wrong answer, or it lied." She was quick to add the technical qualifier — "Of course, it didn't know it was lying because it's not a human being; it's a large language model." The "garbage in, garbage out" framing she applied to AI captures a concern that predates machine learning by decades: a system can only be as reliable as what it absorbed during training, and it has no way to flag its own errors.
Why a Single Failed Query Matters
Atwood is not a technologist, which is exactly why the anecdote travels. Writers, researchers, and historians work in a domain where a confident wrong answer is more damaging than no answer at all. The Father Brown query is the kind of low-stakes factual lookup — a character, a plot point, a broadcast detail — where users expect a clean, verifiable result. It is also the category where AI chatbots most visibly stumble, producing plausible-sounding responses that do not hold up to cross-referencing.
The Commercial Problem Behind the Cultural Moment
For Anthropic and its competitors, Atwood's account illustrates a retention problem that goes beyond engineering benchmarks. A first-time user who tries a product, receives a wrong answer, and publicly describes the experience as unreliable is the kind of story that spreads quickly through the communities — literary, academic, professional — that AI companies most want to convert. A celebrated author trying Claude once, getting a factual error on a British television series, and concluding the tool lied to her is not a crisis, but it is exactly the word-of-mouth that compounds into a perception gap.