Tag

#ELIZAeffect

The ELIZA effect is the name for what happens when people project understanding onto a system that has none, and it is older and more relevant than the chatbot era that rediscovered it. These posts trace the phenomenon from Joseph Weizenbaum's 1966 therapist script to the modern language models that elicit the same projection at industrial scale. I write about why the effect is not a user error but a design outcome, how conversational interfaces are engineered to trigger it, and why Weizenbaum himself spent decades warning that the real danger was not machines thinking but humans abandoning responsibility to systems that seem to. The ELIZA effect is the psychological substrate of every AI deployment that asks you to trust it.