Hinton says AI understands and has emotions; researchers push back

AlexTensor · x · 2026-09-22

Geoffrey Hinton doubled down: AI is not just a tool — "it understands and has emotions." He traces LLMs to a tiny 1985 language model built as a theory of word meaning, arguing their ability to answer tricky questions shows they understand the way humans do, and that skeptics lack a real model of how people work.

Researchers like Philip Torr publicly pushed back: they respect Hinton but see LLMs as nothing beyond a tool — "extraordinary claims should require extraordinary evidence."

Original post →

More from AGI Musings

AGI Musings channel →