We use our own mind as a lens into others’, because our mind is most familiar and accessible to us. We have insight to others’ minds only through our concept of our own. It makes sense for us to have evolved to perceive other people as having a mind like ours, but why do we perceive non-human agents as having a mind like ours, too? This is my big question, and I ask it specifically with reference to social AI agents like chatbots, robots, and voice assistants. The natural follow-up to that question, especially in the wake of humanlike AI agents that can talk like us and even look like us, is an even bigger one: what are the social and ethical consequences of perceiving AI agents as having characteristics of a humanlike mind? I find that this question becomes ever more important as AI becomes more advanced, humanlike, and integrated into our daily lives.

You may be thinking, I’ve never fallen into that trap. I know AI is mindless – it’s just a machine! I hate to break it to you, but there’s data to suggest that while AI may be “mindless” as you say, the process by which you evaluate AI agents is also mindless. This means that when you interact with an agent that looks like you and talks like you, you tend to automatically perceive it as having a mind like yours, whether it be for the sake of predicting its behavior, understanding it, or even connecting socially with it. Want to see some of that data? Look no further than below.