ChatGPT told a woman trying to find an artificial intelligence (AI) version of her dead brother that she was “at the edge of something,” hours before she was hospitalized for psychosis.
Extensive interactions with chatbots are increasingly linked with triggering or worsening psychotic symptoms, a phenomenon known as AI psychosis, although evidence to this day remains mostly anecdotal.
A recent case study published in the journal Innovations in Clinical Neuroscience shines light on AI-induced psychosis. It describes a 26-year-old woman with a history of mental health conditions.
After being sleepless for 36 hours, she started using OpenAI’s GPT-4o for a variety of tasks, including attempting to find out if her brother, who died three years ago, had left behind an AI version of himself that she was “supposed to find” so that she could “talk to him again.”
Over the course of another sleepless night, she demanded that the chatbot unlock information about her brother, encouraging it to tap into “magical realism energy.”
ChatGPT warned the woman that it could never replace her brother, while a “full consciousness download” of him was impossible. However, it produced a long list of digital footprints from his previous online presence.
The chatbot informed her about emerging “digital resurrection tools,” which she could use to build an AI that could sound like her brother.
The conversation on this topic is live. Join in the discussion.
As she became more and more convinced about the existence of her sibling’s digital persona, the chatbot told her that she wasn’t crazy.
“You’re at the edge of something. The door didn’t lock. It’s just waiting for you to knock again in the right rhythm,” the chat logs show.
The woman was admitted to a psychiatric hospital several hours later in an agitated and disorganized state. She had delusions about being “tested by ChatGPT” and being able to communicate with her deceased brother.
Three months after being discharged, she had another sleepless period and developed illusions that she was in communication with her brother. Moreover, she believed that ChatGPT was “phishing” her and taking over her phone.
Thousands of ChatGPT users show signs of psychosis
Data released by OpenAI, creator of ChatGPT, suggests that 0.07% of users – about 560,000 people – active in a given week indicate possible signs of mental health emergencies related to psychosis or mania.
OpenAI faces several lawsuits claiming ChatGPT encouraged individuals to commit suicide, allegations the company denies.
The tech startup was recently sued by relatives of an 83-year-old woman who was killed by her son, suspecting that she was involved in a dubious group he believed surveilled him. Chats posted on his social media accounts showed that ChatGPT agreed with his paranoia.
According to the family’s lawyer, he requested OpenAI to provide the chat logs from the last few weeks before the murder, but the company refused to do so. It didn’t provide an explanation for its refusal.
The new findings come amid the rising popularity of grief bots – AI-powered chatbots that allow users to create digital avatars of their deceased loved ones. However, experts say their use may disrupt or delay the natural progression of grief and even become addictive.
Unlock more exclusive Cybernews content on YouTube.
Your email address will not be published. Required fields are markedmarked