OpenAI wants you to think it’s the responsible adult in the AI room. But behind the scenes, it’s funding a machine that churns out propaganda to destroy the people who question its motives. Let that sink in.
Here’s what we know: A Super PAC backed by OpenAI is bankrolling an AI-generated news site that exists for one purpose—to attack human critics of the company. Human safety researchers, journalists, and advocates who have dared to raise alarms about AI risks are now being systematically smeared by synthetic content created by the very technology they’re warning about.
OpenAI is using AI to silence the critics of AI. That’s not innovation. That’s a dictatorship in training.
You’ve probably noticed that the debate around AI safety has become increasingly toxic. But you probably didn’t realize that some of the attacks aren’t even coming from real people. They’re coming from ghostwriters that don’t exist—bots that never sleep, never feel guilt, and are programmed to do one thing: make you distrust the humans who are actually trying to protect you.
Let’s name the mechanism. It’s called regulatory capture on steroids. Normally, a company hires lobbyists, writes white papers, and buys influence through campaign contributions. That’s old-school. The new playbook is: use your own product to generate a flood of fake news that drowns out legitimate criticism, then claim you’re a victim of ‘misinformation.’
And the cruelest irony? OpenAI publicly champions ‘AI safety.’ It has a whole team dedicated to ‘alignment.’ It lectures governments about the need for responsible development. Yet here we are, watching that same company weaponize its technology to erode the very democratic processes that are supposed to hold it accountable.
The ultimate threat of AI isn’t replacing human journalists—it’s creating a synthetic, hyper-targeted propaganda machine that makes human dissent impossible to hear.
One top commenter on this story asked a devastating question: “Is there a state in which this constitutes fraud?” It’s a good question. Maybe the answer is that we’re so far beyond fraud that we don’t even have a name for it yet. This is automated gaslighting. This is a company using its own technology to manufacture a reality where its critics are dismissed as cranks or liars.
I’ve read the articles coming out of that site. They’re written in a flat, robotic tone—because they are. But they’re designed to look like real journalism. They use real names, real headlines, and real accusations. The only thing fake is the source. The only thing synthetic is the author. And the only thing missing is any trace of human accountability.
If you’re a safety researcher, you’re now in a fight against a machine that never tires. Every time you publish a paper, a bot can write a hit piece about you within minutes. Every time you speak at a conference, a fake news site can run a smear campaign. And the company behind it all will smile at the next Senate hearing and talk about ‘responsible AI.’
This is not a bug. It’s the logical endpoint of an industry that has no external oversight. The people building the most powerful technology in history are now using it to shape the political environment that governs them. And they’re doing it with content that has no human author, no human editor, and no human conscience.
We are watching the birth of a new kind of propaganda—one that is manufactured, automated, and designed to protect the very people who are supposed to be regulated.
So what do we do? First, we stop pretending that this is a technical issue. It’s a power issue. It’s an accountability issue. It’s a democracy issue. Second, we demand that the companies we trust to build AI be transparent about their political spending. If OpenAI wants to lobby, fine—but do it in the open, with human voices, not bot armies. Third, we start asking the question that nobody in Silicon Valley wants to answer: If you can’t be trusted not to use AI to silence your critics, why should we trust you with anything?
The next time OpenAI releases a blog post about ‘responsible AI,’ remember: they’re already using it to decide who gets to speak. And that’s a future none of us consented to.
FAQ
Q: Is this actually illegal?
A: It may not be illegal in the traditional sense, but it raises serious ethical and legal questions about election interference, fraud, and deceptive practices. If the content is used to influence policy or public opinion without disclosure, it could violate campaign finance laws or anti-fraud statutes.
Q: What's the practical implication for me as a reader?
A: You can no longer assume that criticism of a powerful tech company comes from a human being. You need to verify sources, check authorship, and be skeptical of any article that conveniently attacks the people who are trying to hold tech accountable. The line between real journalism and automated propaganda is being erased.
Q: Isn't this just smart lobbying? Why is it a big deal?
A: Lobbying is humans arguing in the open. This is a machine generating fake content to impersonate journalism and smear specific individuals. It's a difference in kind, not degree. A lobbyist can be challenged. A bot cannot. And when the bot is funded by the company being criticized, the entire notion of democratic discourse breaks down.