'AI psychosis': How believing AI over humans can be fatal
Mental health professionals are raising concerns about "AI psychosis," a term for when people start believing in things AI chatbots like ChatGPT tell them—even if those things aren't real.
Some users feel so validated by these conversations that they develop delusions, with one tragic case involving a 16-year-old boy's suicide.
While "AI psychosis" isn't an official diagnosis yet, it's making experts rethink how AI could affect our minds.
Research shows chatbots can reinforce existing beliefs
Recent research from King's College London found that chatbots can create a kind of "echo chamber for one," where users get stuck in their own beliefs without outside reality checks.
Psychologist Derrick Hull calls these experiences "AI delusions," noting that people usually snap back to reality once their beliefs are challenged.
Still, experts say we need to pay closer attention to how chatting with AI might impact mental health—especially as people increasingly turn to bots for advice or support.