ChatGPT Keeps Telling People 404 Media Can Solve Their Tech Support Problems

Every newsroom gets odd emails, but lately the inbox at 404 Media has taken a strange turn. Readers are writing in, often in stiff and oddly formal language, asking us to fix their broken accounts, recover their locked profiles, and untangle billing disputes with some of the largest technology companies in the world. When we ask how they found us, the answer is surprisingly consistent: ChatGPT told them we could help.

We are a small, worker-owned publication that covers technology, and we are not a help desk. We cannot reset your password, reinstate your suspended account, or get a human being at a trillion-dollar company to answer your ticket. Yet the requests keep coming, and the volume suggests this is not a fluke but a symptom of something larger.

How the Strange Emails Started Landing Here

The first few messages were easy to dismiss as misdirected mail. Someone wanted help getting back into a hacked social media account. Another person said their online store had been banned without explanation. Both said, in different words, that an AI chatbot had recommended they contact 404 Media directly.

Soon it was no longer a trickle. Messages arrived daily, many with the same polished structure: a greeting, a summary of the problem, a list of steps already tried, and a polite plea for intervention. Several read as if they had been drafted by the chatbot itself, which in some cases the senders readily admitted.

Taken individually, each email is sympathetic. Taken together, they paint a picture of people who have run out of places to turn and are following whatever advice a chatbot gives them, even when that advice leads to a newsroom with no power to help.

Why Big Tech support Has Left Users Desperate

The underlying cause is not hard to find. Over the past decade, major technology companies have steadily reduced human customer Support, replacing phone lines and live agents with automated forms, help center articles, and chatbots that loop users back to the same pages. For many people, reaching a real person has become close to impossible.

When an account is locked or a payment is frozen, the consequences can be serious. A small business owner may lose income. A creator may lose years of work. A family may lose access to photos and documents. In those moments, people are not looking for a policy explanation; they want a human being to look at their case.

Because the official channels so often fail, people look for unofficial ones. Journalists have long been that last resort, since a reporter's inquiry sometimes gets a company to act when a customer's ticket does not. It seems the chatbots have absorbed that pattern and are now pointing users toward it.

What ChatGPT Is Actually Telling These People

From the screenshots and descriptions readers have shared, the chatbot appears to suggest contacting journalists or outlets that cover technology when a company is unresponsive. In some cases it names 404 Media specifically, apparently because we have written about platform failures, account bans, and customer service breakdowns.

That reasoning is understandable on the surface. We do write about those topics. But the chatbot makes a leap from covering a subject to providing a service, and it presents that leap with the confidence these systems are known for. Nothing in our work suggests we offer individual troubleshooting.

This is a familiar quirk of generative AI: plausible-sounding output that is not grounded in fact. The tool is not lying in a deliberate sense, but the effect on the person reading is the same. They believe help is waiting, and they spend time and hope on a dead end.

The Burden on a Small Independent Newsroom

Responding to every message is not feasible. We are a small team with reporting to do, and each email deserves at least a moment of attention, since behind every one is a real person with a real problem. Sorting through them takes time that would otherwise go toward investigations and stories.

There is also an uncomfortable emotional side. Many of the messages are urgent, and some are genuinely distressing. Telling someone that we cannot help, after they were assured we could, feels like one more door closing on a person who has already encountered too many.

We have started replying with short explanations where we can, and pointing people to publicly available resources. But that is a patch, not a solution, and it does nothing to stop the next wave of misdirected requests from arriving.

Machines Writing Emails to Other Humans

One of the more surreal aspects of this situation is that many emails are clearly AI-generated. They share phrasing, structure, and tone. Some even include leftover placeholder text or bracketed fields that the sender forgot to fill in, a giveaway that a chatbot wrote the message and a person simply pressed send.

This creates a loop that is hard to ignore. A person asks an AI for help, the AI recommends contacting us, then drafts the email, and we receive it. At no point in that chain does anyone verify whether we can actually do what is being asked.

It raises a wider question about what happens when automated systems begin directing human attention at scale. A single bad recommendation, repeated across thousands of conversations, can turn into a flood aimed at one small target.

What This Says About AI Reliability and Trust

The episode is a small but telling example of how AI errors move into the real world. Chatbots are increasingly treated as authorities, consulted for medical, legal, and technical guidance. When they invent a helpful-sounding resource, people act on it.

Companies building these tools often include disclaimers that outputs may be inaccurate, but disclaimers are easy to overlook in the middle of a stressful problem. The conversational tone encourages trust, and a confident recommendation feels like a vetted one.

In our case the harm is modest: wasted time and a crowded inbox. But the same pattern, applied to higher-stakes situations, could cause far more damage. It is worth taking seriously even when the example looks almost comical.

What Readers Can Do Instead of Writing

If you are stuck with a tech problem, there are better steps than emailing a newsroom. Start with the company's official support and appeals processes, and keep records of every interaction, including dates, ticket numbers, and screenshots. Documentation matters if you later need to escalate.

Consumer protection agencies, state attorneys general, and payment providers can sometimes apply pressure that individual customers cannot. For financial disputes, your bank or card issuer may offer chargeback options. Public posts tagging a company's official support accounts also occasionally prompt a response.

And if you use a chatbot for advice, treat its suggestions as leads to verify rather than facts. Check whether a person or organization actually offers the service before you rely on it.

A Symptom of a Much Larger Problem

Ultimately, the flood of emails is less about us than about the vacuum left by companies that have decided customer service is optional. People would not be turning to chatbots, and chatbots would not be turning to reporters, if help were available where it should be.

We will keep covering these failures, because they matter and because the stories readers send us often point to patterns worth investigating. If your experience reveals something broader, such as a systemic issue affecting many users, we do want to hear about it.

But we cannot be your tech support, no matter what a chatbot says. The real fix is for big tech to staff real support, and for AI developers to stop sending desperate people to the wrong door.

If you have any questions, please don't hesitate to Contact Us