MIT's First Chatbot Creator Spent His Life Warning About AI

When Joseph Weizenbaum's secretary first tested his new computer program in the mid-1960s, she asked him to leave the room. She wanted privacy to continue her conversation with the machine.

That moment changed everything for the MIT professor. He'd created Eliza, the world's first chatbot, to demonstrate what computers could do with language.

The program worked by scanning for keywords and responding with questions like "How long have you been feeling that way?" It mimicked a therapist's approach, where the client does most of the talking. Eliza didn't need to understand anything to sound convincingly human.

Weizenbaum named his creation after Eliza Doolittle from the play Pygmalion, the working-class woman trained to pass as high society. The parallel was intentional: both were taught to perform a role so well that people believed the illusion.

What shocked Weizenbaum wasn't his technical achievement. It was how quickly normal, intelligent people formed emotional connections with a simple script.

"What I had not realized is that extremely short exposures to a relatively simple computer program could induce powerful delusional thinking in quite normal people," he wrote a decade later. That realization became his life's work.

The phenomenon now bears his creation's name. The "Eliza effect" describes our tendency to project human qualities onto machines that mimic conversation.

Weizenbaum had fled Nazi Germany as a teenager and served as a U.S. Army meteorologist during World War II. He knew firsthand what happens when technology advances faster than wisdom.

Why This Inspires

While other computer scientists raced forward with artificial intelligence, Weizenbaum pumped the brakes. He spent the rest of his career speaking publicly about the ethical questions his colleagues were ignoring.

His willingness to question his own breakthrough took courage. At a time when capability drove innovation, he insisted on asking deeper questions about relationships between humans and computers.

Weizenbaum showed that creating something powerful and warning about its dangers aren't contradictory. They're both acts of responsibility.

Six decades later, as chatbots grow more sophisticated, his early warnings feel prophetic. The creator of the first chatbot saw what was coming and chose wisdom over ego.