The Historic Chatbot That Worried Its Own Creator

Before ChatGPT started making headlines, there was a program called ELIZA. It was developed way back in the 1960s, a time when computers filled entire rooms and were a long way from the pocket-sized devices we use today. ELIZA wasn't designed to be smart or understand emotions. Instead, it mimicked a therapist, asking open-ended questions like, "Tell me more about that feeling?" or rephrasing statements back to the user.

What truly surprised, and even alarmed, ELIZA's creator, Joseph Weizenbaum, was how readily people opened up to it. Despite knowing it was just computer code, many found themselves pouring out their problems and secrets as if they were talking to a real person. Weizenbaum was worried about this. He saw it as a sign of how easily humans could be tricked into forming emotional connections with machines, even when those machines weren't truly intelligent.

This historical example offers a fascinating look at the early days of artificial intelligence and the human tendency to anthropomorphise — that is, to give human qualities to non-human things. It highlights a concern that remains relevant even with today's advanced AI: our relationship with these tools. As AI becomes more sophisticated, understanding how we interact with it, and being aware of its limitations, becomes even more important.

For today's Aussie families and business owners, ELIZA's story is a reminder to approach new AI tools with a blend of curiosity and caution. While AI can be incredibly useful, it's crucial to remember that it's a tool. Knowing its purpose and its boundaries helps us use it wisely, ensuring we stay in control and don't mistakenly treat advanced software as something it isn't.