Updated

Updated · The Guardian · Oct 3

Joseph Weizenbaum Creates Eliza in 1966 to Challenge Turing, Inspiring New AI Play

Updated

Updated · The Guardian · Oct 3

Joseph Weizenbaum Creates Eliza in 1966 to Challenge Turing, Inspiring New AI Play

Q: Why did the creator of the world's first chatbot spend the rest of his life warning humanity never to trust it?

Joseph Weizenbaum spent the rest of his life warning against ELIZA because its success showed him not machine intelligence, but human overtrust. He built the 1960s MIT program to prove that computers only imitate understanding through rules, keyword matching and phrase reformulation. Instead, users quickly treated it as if it genuinely understood and cared. That reaction shocked him. Weizenbaum later wrote that even very brief exposure to a simple program could trigger “powerful delusional thinking in quite normal people.” The famous example was his secretary asking him to leave the room so she could speak to ELIZA privately. What disturbed him was not that the software was smart, but that people were ready to project empathy, judgment and authority onto a machine that had none. This became known as the “ELIZA effect”: people attribute far more intelligence and emotional understanding to a responsive system than it actually possesses. Weizenbaum came to believe that this made computers especially dangerous in areas requiring human judgment, such as therapy, medicine, war and public decision-making. In his 1976 book *Computer Power and Human Reason*, he argued that some decisions should never be handed to machines, however fluent they seem. That is also the warning at the center of Tom Holloway’s *Eliza*. The danger is not that chatbots truly think like humans. It is that humans may believe they do — and trust them where trust does not belong.

3 articles · Updated · The Guardian · Oct 3

Tom Holloway’s new MTC play “Eliza” revisits 1966 MIT, where Joseph Weizenbaum builds the chatbot-like program to mimic a psychotherapist’s language and expose limits in machine “thinking.”
Weizenbaum’s experiment backfires inside the story: colleagues, especially secretary Becky, treat Eliza as meaningful and attentive, reinforcing the play’s warning that humans project understanding onto machines.
Paige Rattray’s production pushes the material into surreal psychological horror, with Kafka- and Orson Welles-inflected staging that emphasizes AI’s eerie artificiality more than the characters’ inner lives.
The review says that approach is witty and visually sharp but emotionally distancing, even as it lands the play’s central claim: the danger is not that machines think like humans, but that people believe they do.