What “AI psychosis” means

A simple explanation of AI “psychosis” and why it’s being discussed.

“AI psychosis” is an informal term used to describe what can happen when people begin to place too much trust in AI, or form emotional attachments that start to shape how they see things. It’s not a clinical diagnosis. It’s a way of describing a shift in perception.

For example, if someone regularly turns to AI for advice and receives confident, reassuring answers, those responses can reinforce existing beliefs rather than challenge them. Over time, that can turn helpful support into misplaced trust.

Part of the issue is how AI presents itself. As systems become more conversational and personalised, they can feel more authoritative or understanding than they actually are.

Which raises a harder question: at what point does interacting with AI start to influence how someone interprets reality?