ELIZA Effect

The ELIZA effect describes people’s tendency to attribute understanding or empathy to conversational software beyond what its behavior establishes.

Weizenbaum’s 1966 paper describes a program that generates responses through rules for processing and transforming text. Sherry Turkle later discussed users’ emotional engagement and their efforts to sustain the appearance of a responsive companion.

The effect concerns human interpretation. It does not settle whether another system is conscious. Machine Intelligence and AI Alignment pose different questions from whether a conversation feels personal.

Meaningful relationships and attribution

De-humanizing Care investigates how users understand chatbot relationships. Artificial Intelligence in Positive Mental Health discusses oversight and validation. Perceived care, subjective experience attributed to a system, and demonstrated clinical efficacy remain separate questions.