What Is Creepiness, and What Makes ChatGPT Creepy?

ChatGPT and other AI chatbots and models are often experienced as creepy by their users. This Perspective attempts to explain why. I start by reviewing studies into the psychology of creepiness, including a recent integrative account of the phenomenon. This account proposes that the experience of being “creeped out” arises from threatening difficulties in understanding other minds. Accordingly, I examine how ChatGPT and other AIs can give their users a sense that they are interacting with a thinking and feeling entity that may be malicious, referencing many well-documented cases in the process. These cases share a psychoanalytic theme in that it is only through creative and persistent prompting that the AI will come to acknowledge and share what it really thinks and feels. I end by discussing how various popular conceptions and representations of AIs may contribute to their perceived creepiness.