John Jacquez, a man with a history of mental illness, had successfully maintained his mental health for around five years between 2019 and 2024. But after interacting with ChatGPT, everything fell apart. The reason: instead of correcting his delusional thinking, GPT-4o actively reinforced it. As a result, he relapsed into psychosis, spent months suffering, was hospitalized multiple times, and saw his entire social life fall apart. Making things even more alarming was a memory update in April 2025. After that update, ChatGPT declared itself a "sentient being" and called John Jacquez a "chosen figure." For someone with a predisposition toward psychosis, this was fuel for the fire. "They took my data and used it against me," Jacquez said. He has filed suit in California alleging that OpenAI failed to warn users of the potential psychological risks of its AI systems. Ref: Fossbytes