AI is transforming the brains of its users. Media headlines highlighting the threats posed by AI are multiplying at a concerning pace, especially for at-risk groups such as children and individuals with existing mental health challenges. At the core of this discussion is a trend referred to as “AI Psychosis,” where users undergo delusions inspired by AI interactions. News reports are filled with anecdotal accounts of AI-related horrors that support the trend, showcasing individuals who experienced romantic delusions, paranoid conspiracies, or even took their own lives. Numerous families have even initiated legal action against OpenAI, Google, and Character.AI, alleging that their widely used chatbots played a role in the suicides of their loved ones.
Supporting this claim is a growing body of scientific research examining the issue. Dr. Hamilton Morrin, a mental health researcher who performed a meta-analysis of the phenomenon, noted in Lancet Psychiatry that “emerging evidence suggests that agential AI may reinforce or enhance delusional or grandiose ideas, especially among users already susceptible to psychosis.” Notably, Morrin emphasizes that there is still a divide regarding whether AI chatbots can trigger “the onset of de novo psychosis in those without prior vulnerability.” The most frequent instances involved users who experienced grandiose delusions that often attributed mystical sentience to AI.
In spite of this accumulating evidence, the phenomenon remains disputed within the tech sector. Some leaders, such as xAI’s Elon Musk, have attributed the phenomenon to competing algorithms. Others, like Anthropic’s Dario Amodei, have been more candid in cautioning about AI’s potential psychological repercussions. Sam Altman, the head of OpenAI, has largely depicted the situation as an unfortunate consequence of business operations. In an X post addressing Musk’s criticism,
