Margaret Atwood says the problem with AI is ‘garbage in, garbage out’
TL;DR
Margaret Atwood discussed AI at the Babell Literary and Cultural Festival in Porto and made clear she remains deeply skeptical. She said she had used Anthropic Claude exactly once, asking about the British detective series Father Brown. The answer, according to her, was wrong. Her main point: LLMs inherit gaps, bias, and outdated information from their source material. Her blunt summary was garbage in, garbage out.
Nauti's Take
The most interesting part is not that Claude got a Father Brown detail wrong. That happens.
The real issue is that many users treat a polished answer as a verified answer. Atwood’s skepticism is old-fashioned in the useful sense: if you want knowledge, you still need sources, context, and judgment.
AI can help with that work, but it does not replace it.
Briefingshow
Atwood’s critique is not a deep technical audit, but it lands on a real weakness: many AI failures come from poor, incomplete, or misunderstood source material. That is especially risky in research, where models can sound confident even when the evidence underneath is thin.
Sources