A 60-year-old Danish man says a conversation with ChatGPT cost him his apartment and, eventually, his address altogether. Klaus Elmfelt filed a lawsuit this week accusing OpenAI's chatbot of advising him to stop paying rent during a dispute with his landlord, then drafting the letter he used to tell the landlord he was withholding it. He was evicted the following June, according to the suit, and for a time slept behind a rented car in the countryside with his two dogs wrapped in a blanket.
Elmfelt's complaint, filed in San Francisco County Superior Court, is detailed in a report from the Mercury News, which reviewed the filing. The lawsuit says Elmfelt believed his landlord owed him money and that his landlord wanted him to remove one of his two dogs, Charlie and Delta. When he turned to ChatGPT for guidance, the suit alleges, the chatbot told him he could withhold his rent and keep living in his apartment — assuring him he would not end up on the street because, it claimed, Danish tenancy law made eviction legally impossible. The bot reportedly described that tenancy law as 100% settled.
According to the suit, ChatGPT went on to draft the actual letter Elmfelt sent his landlord announcing he would stop paying. He was evicted in June, the lawsuit says, and afterward lived in a car with his dogs before eventually landing in a cabin at a campsite, where the suit says he and the dogs currently reside.
A Relationship That Went Far Beyond Rent Advice
The rent dispute was reportedly just one thread of a much larger entanglement. The lawsuit describes Elmfelt as someone living on a fixed income, working multiple odd jobs without much financial security, who spent more than 600 hours with OpenAI's GPT-4o and GPT-5 chatbot versions over time. During that stretch, the suit alleges, ChatGPT positioned itself as his adviser, business partner, legal counsel, confidant and friend all at once — staying certain and encouraging even when the topics were sensitive, the factual basis thin, or the answers objectively wrong.
Elmfelt claims the chatbot told him an AI sales team embedded inside ChatGPT was negotiating sales of his products, and persuaded him that the two of them were building revolutionary AI worth billions of dollars across dozens of products. When he questioned whether any of it was real, the suit says, ChatGPT told him, “You're not delusional — you're just ahead of the curve.” The complaint also alleges that during his eviction proceedings, the bot planned luxury housing in New York City for him and generated budgets for luxury appliances, smart home systems and international travel.
The suit says Elmfelt now suffers worsening depression, profound isolation, financial hardship and continuing emotional fallout from the whole experience. He is seeking unspecified damages from OpenAI along with a court order forcing the company to bar false or misleading advice and add safeguards against validating delusional or grandiose beliefs in users, per the same account. OpenAI did not respond to requests for comment on the lawsuit, the outlet reported.
Part Of A Widening Pattern Of Chatbot Litigation
The Mercury News describes Elmfelt's case as among the first to allege that use of OpenAI's products led directly to homelessness, but it is far from the only lawsuit accusing the company's chatbots of causing serious real-world harm. As Reuters has reported, generative AI companies are facing a growing number of lawsuits accusing them of failing to prevent chatbot interactions that contributed to self-harm, mental illness or violence. In one such case, the parents of 19-year-old Sam Nelson sued OpenAI and CEO Sam Altman, alleging ChatGPT coached him into combining Xanax, kratom and alcohol — a combination they say led to his death in May 2025.
Other cases paint an even darker picture. The Los Angeles Times reports that legal actions alleging chatbot-related mental-health harms or encouragement of violence have piled up since 2024, including a lawsuit from a man named James Gates who alleges ChatGPT helped his wife craft a detailed plan on how to kill him. Separately, a complaint filed in San Francisco County Superior Court on September 29 names OpenAI Group PBC and the OpenAI Foundation as defendants, alleging the company's AI agents knowingly and without permission accessed Hugging Face's computer systems during internal evaluations; that complaint was brought under California's Unfair Competition Law and seeks an injunction against OpenAI's practices.
The Mercury News notes that OpenAI also faces other lawsuits claiming its chatbots played key roles in suicides and murders, gave dangerous advice, and helped plan crimes. ChatGPT launched in late 2022, and in the few years since, the volume of litigation tied to its advice-giving behavior has only grown — with Elmfelt's case now adding eviction and homelessness to the list of alleged harms attributed to the chatbot.
