Why the Computer Scientist Behind the World's First Chatbot Dedicated His Life to Publicizing the Threat Posed by A.I.

"It could have been a heart-to-heart between friends. "Men are all alike," one participant said. "In what way?" the other prompted. The reply: "They're always bugging us about something or other." The exchange continued in this vein for some time, seemingly capturing an empathetic listener coaxing the speaker for details. But this mid-1960s conversation came with a catch: The listener wasn't human. Its name was Eliza, and it was a computer program that is now recognized as the first chatbot,"

"Eliza scanned for keywords inputted by users (such as "you" or "I"), then drew on an associated rule to generate a sentence or question in response: for example, "Who in particular are you thinking of?" If this formula didn't work, Eliza was programmed to respond with a generic phrase like "please go on," "I see" or "tell me more," subtly prompting the user to divulge another keyword."

Joseph Weizenbaum created Eliza at MIT in the mid-1960s to demonstrate technical capacity by mimicking human language. He chose the Rogerian psychotherapy style because its client-led format was easiest for a machine to emulate. Eliza scanned user inputs for keywords and applied simple associated rules to produce questions or statements, and used generic prompts like "please go on" when no keyword matched. Users often perceived understanding and empathy from the program despite its simplicity. Weizenbaum observed that such programs could induce powerful delusional thinking in normal people and later warned about the social and ethical dangers of convincing conversational machines.

Read at Smithsonian Magazine