AI Discourse Should Remember ELIZA, the 1966 Chatbot That Fooled People Into Seeing a Mind

AlexTensor · x · 2026-09-25

Davetroy argues today's AI debate should recall ELIZA, the early MIT chatbot Joseph Weizenbaum built between 1964-1967. Using pure pattern matching and substitution — its famous DOCTOR script played a Rogerian therapist reflecting users' words back — ELIZA convinced many for decades that there was a ghost in the machine, despite being, in the author's words, dumb as a rock. The point: anthropomorphizing AI output has deep historical precedent.

Original post →

More from AGI Musings

AGI Musings channel →