The Empathy of AI

It correctly identified each of the things that I was struggling with, and then, in its own words, it expressed how I’d been feeling—but far more precisely than I had been able to do on my own. In some strange way, it felt like the AI knew me better than I knew myself.

Dan Shipper

In November of 2022, the company OpenAI offered a linguistically sophisticated AI to the public for the first time in the history of technology. Its name is ChatGPT.

The internet is buzzing with the myrad of use cases for ChatGPT—it’s doing everything from writing legal briefs to writing newsletter articles. The most striking case I found was the use of AI in offering empathetic understanding to real humans.

The quote above is from an article by Dan Shipper (CEO of writer collective “Every”) titled: “Can GPT-3 Explain My Past and Tell My Future?” The author describes his efforts in feeding his personal journal entries to ChatGPT and then asking the AI personal questions about himself based on that data. As the quote suggests, the answers to those questions were so compelling, that he felt that the AI knew him better than he knew himself.

This begs the question: What is empathy, and how can a predictive algorithm run by a machine make us feel known and seen?

The answer might be found in a partner listening exercise from Imago Relationship Therapy. In this exercise, one partner listens to the other by repeating what’s being said word for word—reflecting both the content and emotional tone of the speaker’s language.

Repeating everything you hear as a listener can feel weird at first—it’s kind of like parroting. The trick to the exercise is that both the content and emotional tone are reflected exactly as the speaker means it. No easy task, since we cannot look into someone else’s mind. If the listener is able to do this, though, they have achieved “accurate empathy.” This accurate reflection then opens up a feeling for the speaker of being “seen.”

What’s happening? Perhaps it’s the uncanny experience of witnessing our own thoughts and feelings as an outside observer. It’s kind of like looking up at a mirror. There’s a brief surprise: That’s me; there I am!

Perhaps that’s what the AI achieves in feeding back reflections based on a personal journal. Maybe it’s not the AI who sees us, but something in us that witnesses ourselves.