Margaret Atwood, author of The Handmaid's Tale and The Blind Assassin, didn't hold back when the topic of AI came up at the Babell Literary and Cultural Festival in Porto, Portugal. Her verdict on the technology was characteristically sharp and unsparing.

One Try Was Enough

Atwood revealed she has used an AI chatbot exactly once — Anthropic's Claude — and came away unconvinced. Her test case was straightforward: looking up information about the British detective series Father Brown.

The result was not encouraging.

"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."

This is a distinction that matters to Atwood. The model didn't intend to mislead — it simply has no mechanism for knowing what it doesn't know.

The Core Critique: Garbage In, Garbage Out

Atwood framed the broader problem with AI using a classic computing maxim: "garbage in, garbage out." If the training data is flawed, biased, or low-quality, the outputs will reflect that — regardless of how sophisticated the model appears.

This is not a new critique, but coming from one of the world's most prominent literary voices, it carries cultural weight. Atwood has long been vocal about copyright concerns surrounding AI training data — the question of whose work these models are ingesting, and whether that constitutes theft.

Why This Moment Matters

Atwood's comments land at a time when the creative and publishing industries are deeply unsettled by AI's rapid expansion. Key concerns she and others in literature have raised include:

  • Hallucination — models confidently producing false information
  • Copyright infringement — training on authored works without consent or compensation
  • Quality degradation — an internet increasingly polluted with AI-generated content feeding future models

Her "garbage in, garbage out" framing cuts to a systemic issue the AI industry hasn't resolved: as web content becomes more AI-generated, the training data for next-generation models gets noisier, potentially accelerating quality decay in a feedback loop.

For a writer whose career has been built on precision of language and moral clarity, it's perhaps unsurprising that Atwood finds current AI tools — at least in her one encounter — both unreliable and philosophically troubling.