On Margaret Atwood’s comments at the Babell Literary and Cultural Festival, where the author argued that the problem with AI is “garbage in, garbage out.”
The remark is simple, but it points to a central issue for generative systems: model output depends heavily on the quality, provenance, and assumptions embedded in the data and instructions it receives. Better architecture cannot fully compensate for poor inputs.
Atwood’s criticism also shows how the AI debate has moved beyond technical circles into culture, publishing, and creative labor, where questions of originality and quality are especially visible.