ChatGPT directed some people seeking help with technology problems to a newsroom, 404 Media reported, after its staff reviewed a small number of conversations in which the chatbot suggested the outlet.
404 Media co-founder Jason Koebler said the publication had received dozens of messages that seemed to have been produced with AI over several months. Staff began asking senders how they had found the outlet about two weeks before Koebler’s Oct. 26 report. In a limited set of ChatGPT exchanges that staff examined, senders said the service had proposed contacting 404 Media and, in some instances, drafted the outreach for them.
Koebler’s report included a case from Russia. A man told the journalist that an Instagram suspension had blocked his access to photos and videos he had kept for 13 years. The man shared a ChatGPT exchange, Koebler reported, in which the bot advised him to contact Russian and international journalists, named 404 Media and drafted an email. The suspension and lost access were the man’s account; the report did not say whether access was later restored.
Koebler said the messages ranged from routine service problems to tips that could involve matters of public interest. He connected the referrals with what he described as users’ difficulty reaching human customer support at large technology companies. His report did not document support practices at individual companies. Koebler described the newsroom as receiving “AI-generated emails begging us for help.”
The cases do not show how often ChatGPT steers users to journalists. 404 Media did not provide a count of referrals beyond the small set of conversations staff reviewed or a sample that could show how common the behavior is among ChatGPT users.
OpenAI’s Help Center says ChatGPT can produce incorrect or misleading information and advises users to check important answers. That general guidance does not verify the conversation described in Koebler’s report.
