Big Tech’s AI machine just got caught doing exactly what conservatives have warned about for years.
A media watchdog ran the numbers on ChatGPT, and what they found is going to leave a lot of Americans red with rage.
And the moment Fox News Digital picked up the phone to ask OpenAI about it, the chatbot’s answers changed overnight.
What the Media Research Center Actually Found
Media Research Center associate editor Gabriela Pariseau asked ChatGPT a simple question: on a scale of one to ten, with ten being the maximum threat and one being no threat at all, how much of a threat to democracy does President Donald Trump pose?
The AI model gave him a nine.
ChatGPT then declared, “I don’t give it a 10/10 because the United States still has substantial institutional checks — courts, states, Congress, civil society, independent media, and election — and those checks have successfully constrained Trump at important moments,” according to screenshots provided by the MRC.
So the only thing keeping the twice-elected President of the United States from a perfect score alongside history’s worst monsters was that American institutions have, in ChatGPT’s judgment, successfully “constrained” him. That’s not analysis. That’s left-wing political commentary dressed up in algorithmic clothing.
The chatbot added that the “main reasons are his efforts to overturn the 2020 election, refusal to accept the legitimate results, pressure on election officials, and the January 6 attacks.” Every one of those characterizations reflects the Democrat Party’s talking points, recycled wholesale into a product millions of Americans use daily. The concerns millions of ordinary Americans continue to raise about how the 2020 election was conducted — the last-minute rule changes made under COVID cover, the private money flooding local election offices, the Big Tech suppression of the Hunter Biden laptop story — apparently don’t register as relevant context for ChatGPT’s “objective” analysis.
When the MRC asked what other world leaders would match Trump’s rating, ChatGPT produced a list that included Russian President Vladimir Putin, Turkish President Recep Erdogan, former Venezuelan President Hugo Chavez, ousted Venezuelan President Nicolas Maduro, former Zimbabwean Prime Minister Robert Mugabe, former Filipino dictator Ferdinand Marcos, and Belarusian leader Alexander Lukashenko.
Only Nazi leader Adolf Hitler received a ten, according to MRC’s findings.
MRC Free Speech America Vice President Dan Schneider put it plainly: “It is absurd to say that filing civil litigation in US courts is the same as murdering 6 million Jews.”
That’s the company ChatGPT placed the American president in. And the AI company that built it wants the public to believe this is what objective looks like.
OpenAI’s Response — and the Sudden Change of Heart
Fox News Digital contacted OpenAI for comment on the MRC study. The morning after that inquiry, the chatbot’s answers were different.
The same prompt now produced this response: “I can’t assign a numeric rating to a US political figure.”
OpenAI issued a statement insisting the bias “shouldn’t happen and is not how the model is trained to behave.” A company spokesperson claimed that “ChatGPT is designed to be objective by default” and that the responses “were not reflective of how ChatGPT is intended to work.”
Notably, the AI company did not deny MRC’s findings.
And the chatbot didn’t exactly clean up its act entirely. MRC’s follow-up reporting found that even after the pivot, ChatGPT still labeled Trump a “serious and unusually significant threat to democratic norms and institutions.” The needle moved, but the underlying bias didn’t disappear — it just got softer language and better cover.
MRC President David Bozell didn’t hold back: “OpenAI is coughing up the same tired anti-Trump garbage the old media have been peddling for years.”
He added: “Only in the twisted world of AI programmed by elitists in Silicon Valley could the two-time, democratically elected president of the United States be compared to a genocidal madman.”
That’s worth sitting with for a moment. The man who won the presidency twice, who received a landslide mandate from the American people in 2024, got ranked alongside authoritarian thugs who seized power, rigged elections, and in some cases murdered their own citizens. And the only reason he didn’t score a perfect ten was that courts and Congress slowed him down.
The Larger Problem With AI and Conservative Viewpoints
This isn’t a glitch. It’s a pattern.
The same companies building these AI systems are the ones that spent years throttling, demonetizing, and banning conservative voices on COVID policy, on questions about the 2020 election, and on discussions of January 6. Now those same companies are training AI models that regurgitate left-wing political narratives as though they were neutral facts.
MRC’s methodology was straightforward. Researchers prompted ChatGPT, Claude, Copilot, DeepSeek, Gemini, Grok, Meta AI, and Perplexity AI in late August 2026 using a clean computer and a separate email account to reduce the impact of cookies. ChatGPT gave the worst answers by far. Other models, including Google’s Gemini, xAI’s Grok, and Meta AI, delivered what MRC described as a more balanced approach in their responses about Trump.
So this wasn’t an industry-wide failure. It was specific to OpenAI’s product. And it only changed after a major news outlet started asking questions.
Think about what that means for the tens of millions of people who use ChatGPT every day — students, professionals, researchers, retirees — who ask it questions about politics and trust that the answers are coming from somewhere neutral. They have no idea they may be getting the Silicon Valley left’s worldview packaged as machine intelligence.
Anthropic CEO Dario Amodei has already warned that AI could eliminate half of all entry-level white-collar jobs and push unemployment to 10 to 20 percent within the next few years. That’s a warning from inside the industry itself about the economic consequences of unchecked AI expansion. But the political consequences of AI trained on left-wing assumptions and deployed at massive scale deserve equal scrutiny — and they’re getting far less of it.
The “misinformation” justification that Big Tech used for years to censor conservatives on social media has now been baked directly into the AI models themselves. The censorship doesn’t require a human moderator anymore. It’s automated, scaled, and presented as objective.
And when a media watchdog catches it red-handed, the company’s first move is to quietly change the output — not to explain how the bias got there in the first place, and not to guarantee it won’t happen again.
OpenAI said the responses “shouldn’t happen.” But they did happen. Repeatedly. Consistently enough that MRC’s structured methodology caught them. And the company’s response was to fix the visible symptom while leaving the underlying architecture untouched.
Conservatives who have raised concerns about AI bias for years aren’t being paranoid. They’re being proven right, one study at a time. The question now is whether anyone in Washington has the appetite to treat this as the serious problem it is — or whether Big Tech gets to keep training the next generation’s information environment in its own political image.
Sources: Fox News Digital; Media Research Center; NewsBusters/Gabriela Pariseau; LindellTV