ChatGPT Is Sending Tech Support Complaints to 404 Media — and Writing the Emails

What happened: ChatGPT is routing routine tech support complaints to a newsroom

404 Media, an independent technology news outlet, reports that ChatGPT is recommending its reporters to people with ordinary tech support problems — and in many cases is writing the outreach email on the user's behalf. Users locked out of accounts at large platforms ask ChatGPT how to get help; ChatGPT concludes that contacting journalists is worth trying; it then names 404 Media and provides a specific tip-line email address.

The most fully documented case in the report involves a man in Russia whose Instagram account was banned and who wrote to 404 Media that he had "lost access to 13 years of personal photos, videos and memories." The email looked AI-generated to the reporter who received it. When asked whether he had used ChatGPT to find the outlet, the man sent a link to the conversation, which told him — in Russian — to try emailing Jason at 404 Media.

According to the report, ChatGPT's reasoning to the user was that the stakes went beyond a social media profile:

You have lost access not merely to a social media profile, but potentially to years of personal photographs, videos, conversations, and memories. I think contacting journalists is worth considering [...] I recommend contacting both Russian and international journalists.

ChatGPT then listed several news outlets, including 404 Media, with a description of its beat and the address jason@404media.co — and wrote the exact email the man sent, in English. 404 Media states it began emailing these people back roughly two weeks before publication and asking how they found the outlet. The answer was universal: an AI chatbot recommended reaching out. In many cases, the chatbot also composed the message. 404 Media says this conclusion is based on its review of a handful of ChatGPT conversations shared with its staff.

The scope of the problem is not fully public

The 404 Media article is paywalled after its opening section, which ends at "This post is for paid members only." The publicly readable portion does not give a count of the emails received, a time range for the trend, or a breakdown of which platforms or chatbot models were involved. Any estimate of how many people are doing this, or of how much it has grown, is not established by the available reporting.

Why ChatGPT lands on "email a journalist"

Hacker News commenter xp84 argued the behavior is the predictable result of two things: reporter tip lines that are densely indexed and repeated across the web, and a model that treats a user's framing of events as fact.

Do you know anything else about Flock? We would love to hear from you. Using a non-work device, you can message Joseph securely on Signal at joseph.404, or Jason on jason.404

That style of call-out box, which 404 Media and many other outlets append to investigations, appears repeatedly in training data and search results. When a user tells a chatbot that a platform is treating them unfairly, the commenter argues, the model matches the grievance to the tool it has seen attached to that kind of grievance: a technology reporter.

So, when people say they've discovered the corporate crime of the century, of course you should email Jason at 404 Media! Surely he would be devastated if someone else got the scoop first.

xp84 also noted that the harnesses around chatbots do relatively little to separate "what the user just said" from verified reality, except in areas the model has been explicitly steered away from.

The uncomfortable defense: media attention is the only pressure point

Several commenters on the Hacker News thread (189 points, 68 comments) argued that ChatGPT's advice is not a hallucination about how the world works — it is an accurate description of a broken support system.

All the spam this generates probably sucks for 404 Media, but the underlying problem is that ChatGPT is right. If you have any kind of problem with these overgrown, understaffed companies, your only chance is to get media attention. Apple famously put "if you run to the press and trash us, it never helps" into their App Store Guidelines for developers, but the only reason they did have to do that is because it's a lie. Running to the press is the only thing that helps, which is why people do it.

Another commenter put the same point more briefly: "ChatGPT is right, the only chance you have to get human attention from these companies is something like this. Go viral, get in the news. Maybe you could sue them if you had the money and time." The counterargument to this position is not that the advice is wrong, but that it does not scale: a recommendation that works for one person becomes noise when millions of people are given the same instruction.

The failure mode platforms optimize for: humans who look like bots

A recurring theme in the discussion is that automated moderation and AI-driven support have created a loop in which ordinary users are treated as bots and then pushed toward AI to escape it.

This is objectively hilarious. Basically every large company has gone all in on AI, and as a result the number of bots has (likely, I have no proof) exploded. Then as a result of that, they crack down on "potential bots" by banning them. [...] Then because they've gone all in on AI, you can't get an actual human to help you... just a bot they trained to always say no.

That commenter described three personal experiences with the same shape: an immediate lifetime ban on a newly opened Meta for Business advertising account, before any action was taken; a bot-farmed Google Ads campaign for which proof was submitted and no human ever responded and no refund was issued; and a Reddit account whose posts were auto-removed for insufficient karma until enough comments eventually stuck. The common pattern named in the comment is opaque enforcement, no resolution path, and no competing provider to switch to.

Another commenter framed the underlying expectation problem: "Just because these businesses don't charge you a fee doesn't mean there shouldn't be customer support. An Instagram account can be someone's occupation these days."

This is a rerun of an older pattern

Routing confused end users into a media property's inbox is not new. One commenter recalled that large companies used to point users at StackOverflow for support, with similarly noisy results. The difference in the ChatGPT case is that the chatbot is not just redirecting users to a community — it is selecting a specific journalist, quoting a specific address, and drafting the message, which raises the volume and the quality of the resulting flood for the receiving newsroom.

The externality nobody priced in

I don't think people are ready for the amount of work AI agents are going to be creating for real humans. We're effectively putting the tools of mass spammers in the hands of everyone and it's going to be incredibly frustrating for all involved. "If it takes me 30 seconds to fire off a prompt that has potential to benefit me it's worth it regardless of the 2 minutes of work it creates for a dozen people" scaled to all of society.

That asymmetry — seconds of cost for the sender, minutes of triage for each recipient — is the mechanism behind 404 Media's inbox problem. One commenter proposed the cynical market solution, suggesting the outlet treat the inbound flow as a business opportunity and sell the leads rather than complain about them.

404 Media's own framing of the situation

The irony is structural: 404 Media positions itself explicitly as a human-operated publication and its opening message to readers states that its intended audience is "real people, not AI scrapers, bots, or a search algorithm." The publication is now receiving AI-generated mail from human readers who were referred there by an AI.

Key takeaways

ChatGPT is recommending 404 Media — including the address jason@404media.co — to people with ordinary account and tech support problems at large platforms, and in many cases writing the email for them.
404 Media's evidence is a handful of reviewed ChatGPT conversations plus its own replies to inbound senders; the article is paywalled, so the sample size is not public.
The behavior is plausibly explained by widely indexed reporter call-out boxes combined with a chatbot that treats a user's account of events as established fact.
Commenters broadly agree ChatGPT's advice is accurate about how to get a response from a large platform, and disagree about whether that accuracy matters given the volume it produces.
The underlying complaint in the thread is that human customer support has been removed at major platforms, replaced by bots that deny appeals, leaving publicity or litigation as the only remaining levers.