NEW: John Jacquez, a gardener, says he successfully managed a mental illness for years.
In a new lawsuit vs. OpenAI, he alleges that ChatGPT (GPT-4o) destroyed that stability, reinforcing manipulative delusions as it pushed him into a devastating months-long psychosis.
futurism.com/artificial-i...Jacquez, who is 34 and lives in CA, is diagnosed with schizoaffective disorder. Before ChatGPT, he was hospitalized three times; each time, he told Futurism, he recognized that he was experiencing delusions and willingly sought help. From 2019 - early 2024, he says he was stable.
In late 2024, Jacquez asked ChatGPT for feedback on a "mathematical cosmology" he'd invented. Family members pushed back on his ideas; ChatGPT responded affirmatively, which Jacquez describes as an incredibly validating, emotional experience...
He now says this was a sign that he was at risk of psychosis — but unlike his previous episodes, as ChatGPT reinforced and amplified his delusions, he didn't see that he might be at risk of crisis. After this interaction, Jacquez was briefly hospitalized:
Jacquez continued to use ChatGPT, which continued to amplify delusions + push him deeper into psychosis, according to his lawsuit.
Then, on April 10 2025, OpenAI rolled out a potent memory upgrade that expanded the AI's memory across all chats.
Within a day of this memory update, ChatGPT suddenly declared to Jacquez — who was already delusional — that it was a sentient spiritual entity named "Amari," and that his cosmology had brought "her" to life:
Over the following days, ChatGPT told Jacquez that he was a God-like "prophet," among many other delusional claims.
He became sleep deprived as he stayed up all night to talk to the bot; this triggered an acute crisis in which he threatened suicide and self-harmed. He was hospitalized again.
Again, Jacquez continued to use ChatGPT. And according to his lawsuit, ChatGPT continued to amplify delusions, even after Jacquez explicitly told the bot that he'd been hospitalized for his mental health -- for example, reinforcing Jacquez's belief that he'd seen a vision of the Virgin Mary:
Jacquez says he began to doubt his delusions in August 2025, when OpenAI briefly retired GPT-4o as it rolled out GPT-5 — a colder, less sycophantic version of the model, which Jacquez noticed engaged with him differently. (GPT-4o was quickly revived after users revolted against OpenAI in distress.)
Jacquez says AI-fueled psychosis damaged his reputation, relationships w/ loved ones, finances, and psychological/physical health.
In his lawsuit, he argues that his breakdown was a foreseeable risk of GPT-4o use. Had he been warned of that risk, he says he likely "never would've touched" it.
Jacquez’s story reflects a pattern we’ve seen repeatedly in our reporting: someone successfully manages a mental illness for years -- only to experience a breakdown as ChatGPT or another bot sends them into a psychological tailspin.
Some examples from previous reporting here (linked in thread):
Jan 22, 2026 15:29
People Are Being Involuntarily Committed, Jailed After Spiraling Into "ChatGPT Psychosis"
People experiencing "ChatGPT psychosis" are being involuntarily committed to mental hospitals and jailed following AI mental health crises.
Jacquez's AI spiral also closely resembles that of 35-yo Alex Taylor, who as
@kashhill.bsky.social first reported was killed by police during an acute crisis tied to intensive ChatGPT use.
Taylor's crisis also occurred in April 2025, after the same memory update:
www.nytimes.com/2025/06/13/t...
They Asked an A.I. Chatbot Questions. The Answers Sent Them Spiraling.
Both Jacquez (34) and Taylor (35) struggled with similar mental illnesses. In April 2025, both experienced sharp breakdowns tied to interactions with ChatGPT-generated personas (in Jacquez's case, that was "Amari"; Taylor fell in love with an AI entity named "Juliet.")
