I recently recorded a podcast with my co-host, Lilian Borges, in which we discussed the utility of artificial intelligence in our profession. I have to confess that I needed to do a lot of research; I have used ChatGPT only occasionally—and in fact, I sometimes even forget it exists. Researching for the podcast challenged and interested me in looking into how it’s currently being used in my field.

Some of you may be old enough to remember ELIZA, an early natural language conversation program developed in what would become MIT’s AI lab. Its creator, Joseph Weizenbaum, was astonished years later to learn that people attributed human thought and feelings to it: "I had not realized,” he wrote, “that extremely short exposures to a relatively simple computer program could induce powerful delusional thinking in quite normal people." Some therapists used ELIZA (which was programmed as a Rogerian practitioner so it could echo back what the person said to it) as a prescreening tool. Some went even further and used it in conjunction with their own work.

Zoom up to 2025, and ChatGPT is light-years ahead of its ancestor—inspiring users to believe it is fully like them, with emotions and preferences and empathy. And that, in turn, makes some people ask the obvious question: If I can discuss my problems with a chatbot for free, why pay a therapist?

I understand that. In some cases, a chatbot may be........

© Psychology Today