‘AI psychosis’: could chatbots sustain delusional reasoning?– podcast|Scientific research

There are increasing reports of individuals experiencing deceptions after extensive use of AI chatbots. The phenomenon, referred to as ‘AI psychosis’, has elevated worries that attributes built into large language models may add to some individuals losing touch with reality. Madeleine Finlay speaks with Dr Hamilton Morrin, a psychiatrist and scientist at King’s College London, concerning his current preprint exploring who is at risk and exactly how models might be made safer