Rod Pardo

In the 1960s, an MIT scientist built the first chatbot. It pretended to be a therapist. It didn't understand a single word you said. It just kept asking you to tell it more. His own secretary asked him to leave the room. His name was Joseph Weizenbaum. The program was called ELIZA. He published it in 1966 and named it after Eliza Doolittle, the flower girl in Pygmalion who is taught to speak like a lady. Type that your boyfriend made you come here, and it answered: "YOUR BOYFRIEND MADE YOU COME HERE" That was the whole trick. His secretary had watched him build it for months. She knew it was only code. After a few sentences she asked him: "Would you mind leaving the room, please?" She wasn't the only one. He wrote that some people were very hard to convince they weren't talking to a human. When he suggested he might look at the conversations people had with it, he was "bombarded with accusations" that it "amounted to spying on people's most intimate thoughts." Then it got worse. Practicing psychiatrists seriously believed a program like this could grow into "a nearly completely automatic form of psychotherapy." Weizenbaum was shocked. In 1976 he published a book warning people about computers. He wrote that a very short time with a simple program could cause "powerful delusional thinking in quite normal people." He spent the rest of his life warning the world about what he had built. He had fled Nazi Germany with his family as a boy. In 1996 he moved back to Berlin. He died in 2008, at 85. Sixty years later, telling a chatbot your problems is normal. He never built a machine that understood people. He built the first proof that it doesn't have to. Have you ever told a chatbot something you've never told a person?