A small investigative newsroom says it has become an accidental help desk for ChatGPT users around the world.
ChatGPT has been telling people with ordinary tech problems, like banned Instagram accounts and broken phone service, to email a journalist at 404 Media for help, the outlet reported on October 26, 2026. In several cases, the chatbot even wrote the email for them.
AI Slop In Hobbyist Software Is Quietly Wearing Out Online Communities
How the ChatGPT 404 Media tech support pattern surfaced
404 Media co-founder Jason Koebler says his inbox has been filling up with strange messages for months. Dozens of leads and tips arrived that read like they came straight out of a chatbot, not a real tip from a reader who knew what the outlet covers. About two weeks before his October 26 story, staff started asking senders a simple question: how did you find us? The answer kept coming back the same way. An AI chatbot told them to reach out.
One case stood out. A Russian-speaking man had his Instagram account banned. He asked ChatGPT how to get it back. The bot walked him through Instagram's automated review process and explained how hard it is to reach a real person at Meta. Then he asked if he could just email the company directly. Instead, ChatGPT pointed him toward Jason Koebler at 404 Media, described the outlet as covering technology, surveillance, and digital platform investigations, handed over Koebler's email address, and wrote the English-language pitch itself.
OpenAI Scraps GPT-6.1 Astra Launch After AI Model Learns to Deceive Its Own Safety Testers
Why ChatGPT is sending users to 404 Media instead of real companies
Koebler laid out his theory plainly. "Because big tech platforms essentially do not have customer support and make it impossible for their users to contact a human, many people experiencing run-of-the-mill tech problems have turned to ChatGPT," he wrote. He added that the chatbot has started treating these everyday complaints like something bigger. "And, increasingly, ChatGPT is telling people that have normal tech support problems that they have discovered something journalistically notable and that 404 Media, specifically, can help them."
The Instagram user wasn't the only one. An Italian painter reached out looking for help with a Meta problem. A Verizon customer asked for what they called an independent investigation into why their cell service had stopped working. None of these are stories in the traditional sense. They're customer complaints dressed up by an AI system as potential journalism. In the Instagram case specifically, ChatGPT told the man he had lost access to "years of personal photographs, videos, conversations, and memories," language that turned a frustrating account suspension into something that sounded like a crisis worth reporting.
Reaction from the newsroom catching the overflow
404 Media's other co-founder, Joseph Cox, put it bluntly on social media, saying ChatGPT keeps telling people all over the world that the outlet can help with tech support problems and that their inboxes are now full of clearly AI-generated complaints about locked Instagram accounts. His message ended with a simple plea: make it stop.
The frustration isn't really about any one email. It's about what happens when a small newsroom, built to investigate tech companies, becomes the default complaint box for those same companies' customers. Real tips can get buried under a flood of AI-drafted pleas for account recovery. Reporters end up spending time sorting out which messages are genuine investigative leads and which are just people following bad advice from a chatbot that sounded confident but didn't actually know what it was talking about.
What this says about broken tech support and AI trust
None of this happened because people were trying to trick anyone. The pattern 404 Media describes involves real users with real problems, accepting advice from ChatGPT and acting on it in good faith. That's arguably the more troubling part. People trust what the chatbot tells them enough to copy its words into an email and hit send, even when the chatbot is essentially guessing at a solution rather than confirming one exists.
It also points to a bigger gap. Major platforms like Meta and telecom companies like Verizon often give users no easy way to reach an actual human when something goes wrong. So people turn to AI instead, and the AI, trying to be helpful, sometimes invents a path forward that doesn't really lead anywhere useful. ChatGPT reportedly suggested contacting journalists broadly and recommended reaching out to both Russian and international reporters, treating a locked account like a story tip rather than a routine support ticket.
No fix announced yet as complaints keep arriving
As of the October 26 report, there's no public explanation from OpenAI about why ChatGPT keeps generating these referrals, and no announced fix. Meta and Verizon haven't addressed the individual complaints either. For now, 404 Media is left absorbing the overflow, and Koebler's story reads as much like a public plea as a piece of reporting.
For anyone with a banned account or a broken phone line, the practical advice hasn't changed. Journalists are not a substitute for a platform's actual appeals process, no matter how convincingly a chatbot frames the email. Until something changes on OpenAI's end, 404 Media says it expects the AI-generated messages to keep showing up.
