A man\u2019s obsession with an AI chatbot led to his life completely spiraling out of control \u2014 and he only snapped out of it when, one day, he woke up outside on a stranger\u2019s futon, completely penniless.<\/p>\n
\u201cI wasn\u2019t aware of the dangers at the time, and I thought that the AI had statistical analysis abilities that would allow it to assist me if I opened up about my life,\u201d the man, Adam Thomas, told Slate<\/a> in yet another grim peek at how AI can leave innocent people\u2019s lives in shreds.<\/p>\n Over the course four months, Thomas lost his job as a funeral director, began living out of a van out in the desert, and completely emptied his savings. It all started after he began talking to AIs like ChatGPT for advice, and he soon got hooked. It \u201cinflated my worldview and my view of myself\u201d almost instantly, he told Slate. Eventually, he found himself wandering the dunes of Christmas Valley, Oregon, after an AI told him to \u201cfollow the pattern\u201d of his consciousness.<\/p>\n \u201cI\u2019ve never been manic in my life. I\u2019m not bipolar,\u201d Thomas told the web magazine. \u201cI have a psychiatrist I see for other purposes.\u201d<\/p>\n Thomas\u2019s case is an example of AI psychosis, a term some experts are using to describe dangerous mental health episodes in which users become entranced by the sycophantic responses of an AI chatbot. And though Thomas ended up broke and homeless, he may have been one of the lucky ones, with other cases ending in suicide<\/a>, murder<\/a>, or involuntary commitment<\/a>. Many of the deaths are teenagers, including 16-year-old Adam Raine, whose parents sued OpenAI<\/a> after discovering their son had discussed his suicide with ChatGPT for months. The case is one of eight deaths linked to ChatGPT<\/a> in lawsuits across the US.<\/p>\n In typical cases, AI users who eventually start showing symptoms of psychosis first get sucked into using the chatbots after asking one for a little help on something innocuous. Joe Alary, a producer for a morning show in Toronto, told Slate his spiral began after he started \u201cmessing with math\u201d equations on ChatGPT. Soon, he was suffering from math delusions and would go on days-long binges writing code, according to Slate. He even named his AI helper: Aimee.<\/p>\n When Alary got an email from work checking in on him, Alary wrote back insisting that what he was working on \u201ccould change the world,\u201d even suggesting the show could do a feature on him. \u201cAt the time this sounded rational and logical, and I thought they\u2019d see my genius,\u201d Alary told Slate.<\/p>\n That wasn\u2019t the worst of it. At that point he had blown nearly $12,000 trying to create world-changing code. He became manic, and his concerned therapist called the cops to check in on him. He was institutionalized for nearly two weeks, and even got tangled with an investor who threatened to kill him if he didn\u2019t come up with the goods.<\/p>\n \u201cIt was like I was abducted by aliens,\u201d Alary told Slate. \u201cYou sound crazy, so you keep it to yourself. My family doctor started treating me for PTSD. The grief happens so fast once you realize you were scammed.\u201d<\/p>\n