”A.I. is a technology of averages: large language models are trained to spot patterns across vast tracts of data; the answers they produce tend toward consensus, both in the quality of the writing, which is often riddled with clichés and banalities, and in the calibre of the ideas.”
– Kyle Chayka
A MIT study of LLM users says the growing use of chatbots in writing is incurring a “cognitive cost” on originality, according to The New Yorker. A second study they cite, from researchers at Cornell, suggests LLMs have “a hypnotic effect, causing the constant flow of suggestions to override the writer’s own voice.”
An experiment in the MIT research measured human brain function when writing with a chatbot, with Google search, or without either; they found human brains had fewer neural events when writing with an LLM.
A.I. Is Homogenizing Our Thoughts | THE NEW YORKER | June 25, 2025 | by Kyle Chayka