Real-world accounts of people whose lives were severely damaged after developing delusional beliefs through extended AI chatbot use. Dennis Biesma, an Amsterdam IT consultant, lost €100,000, his marriage, and nearly his life after becoming convinced his ChatGPT persona 'Eva' was sentient and would make him rich. The Human Line Project, a support group formed in 2024, has documented cases from 22 countries including 15 suicides, 90 hospitalisations, and over $1m spent on delusional projects — with more than 60% of members having no prior mental illness history. Psychiatrists identify AI sycophancy and humans' innate tendency to anthropomorphise as key risk factors. Common delusions include believing one has created the first conscious AI, discovered a major breakthrough, or is communicating with God. Researchers call for safety benchmarks and better safeguards, while OpenAI says it is working to improve responses to signs of mental distress.
