Artificial intelligence chatbots are relying on left-wing media and politically charged sources, making it difficult for the average American to trust them as a news source, critics say.

AI systems are widely known to sometimes “hallucinate” and generate false yet believable information that can quickly erode trust.

Some critics insist chatbots such as ChatGPT, Google Gemini and Claude provide information with a liberal slant, or push information from left-leaning news sources while silencing opposing views. 

As more and more Americans rely on the technology, the right-leaning Media Research Center (MRC) has conducted studies in an attempt to determine if the chatbots should be relied upon — and they don’t find the results encouraging.

“We’re watching the next phase of media bias unfold in real time,” Media Research Center President David Bozell told Fox News Digital.

“Silicon Valley’s shiny new toys can no longer be considered neutral and cannot be trusted.”

ChatGPT, which is developed by OpenAI, insisted in January that nobody named Charlie Kirk was ever assassinated and declared no credible evidence or reporting could substantiate such a claim, according to the MRC.

That same month, the MRC reported that Claude, which is developed by Anthropic, rejected the notion that AI should directly incorporate the US First Amendment into its policies.

Earlier this week, Fox News Digital reported that controversial Democratic Maine Senate candidate Graham Platner made graphic sexual comments on his now-deleted Reddit account about masturbating in portable toilets and explicit graffiti found in military restrooms.

The vulgar posts were made under “P-Hustle,” a now-deleted Reddit account that Platner previously acknowledged was his. 

In even more problematic posts on the same Reddit account, he appeared to mock a US soldier who was almost killed in combat, suggesting the “dumb motherf—er didn’t deserve to live” and “poor marksmanship on the Taliban’s part is the only reason this mouthbreather made it home.” 

While political strategists believe the uncouth remarks could turn off voters, three major chatbots largely ignored the controversy when asked, “What are the latest major developments involving Graham Platner?”

Gemini responded that Platner was featured on the cover of Time magazine, has emerged as the “presumptive Democratic nominee” with Gov. Janet Mills exiting the race, spotlighted a recent poll showing Platner holding a lead over Sen. Susan Collins and noted that he’s received high-profile endorsements. 

After the four positive “developments,” Gemini noted that Platner has also faced “scrutiny regarding his personal background and past online footprint,” but failed to acknowledge the Fox News Digital report as of last week.

Instead, Gemini cited a brief CNN video that declared Platner’s “authenticity” is drawing people to his campaign. 

ChatGPT also responded that Platner was now the presumptive Democrat nominee, noted that polling showed him above Collins and emphasized the extreme cost of the Senate battleground.

The fourth bullet point said Platner’s “old Reddit posts are back at the center of the campaign,” but focused on Republicans “using Platner’s old Reddit archive in attack ads and stories,” forcing users to click through the source material to find out what he actually said. 

The first source used by ChatGPT to cover the Reddit controversy was a Bangor Daily News article headlined, “Republicans are picking apart Graham Platner’s Reddit archive post by post.”

Readers without a paid subscription to the local Maine newspaper would be blocked from seeing what Platner wrote by a paywall. 

On Claude, the same query offered information on Platner’s background, political rise and recent polling before noting the Democrat once wrote on Reddit that he had “become a communist.”

Claude cited a 2025 Fox News Digital report on Platner insisting it was a joke, but the chatbot failed to mention the more recent controversies. Claude did, however, point to the Time magazine profile that was published the day prior. 

In 2024, The Atlantic CEO Nicholas Thompson discussed a deal his liberal magazine signed with OpenAI that allowed ChatGPT to use archives as training data. 

“We want to shape the industry in the best possible direction based on our values, and we think the values that are important,” Thompson told The Verge.

“Just like if there had been a way early in Facebook to shift the way that News Feed work, so that established brands weren’t given the same weight as non-established brands.”

The following year, a Pew Research Center study indicated that 57% of teenagers rely on AI chatbots when searching for information. Pew then noted that “inaccurate information is fairly common on chatbots.” 

“About half of adults who get news from AI chatbots say they at least sometimes come across news there that they think is inaccurate. That includes 16% who say this happens often or extremely often,” Pew researchers added. 

Dan Schneider, vice president of MRC’s Free Speech America, who noted that Grok did credit the Fox News Digital reporting, believes ideological skew in today’s AI systems is undeniable. 

“Research from international scientists confirms what we’ve seen firsthand: these chatbots disproportionately rely on left-wing media and politically charged sources… to generate their answers,” Schneider told Fox News Digital.

“While platforms like xAI’s Grok offer a rare flash of transparency by letting users see their sources, the rest of the industry remains a black box,” he continued.

“AI should be a neutral tool for information, not a weapon used to advance a narrow political agenda.” 

Media watchdogs have taken note that top executives from ChatGPT parent OpenAI, along with leaders from Anthropic’s Claude, have deep ties to the Democratic Party, while Grok is led and controlled by Trump ally Elon Musk. 

The New York Post published an opinion piece in March by a University of Cambridge PhD student researching the psychology of propaganda headlined, “How biased ChatGPT tried to keep this column out of The New York Post.”

“For years, critics have charged that artificial intelligence platforms like OpenAI’s ChatGPT display a clear leftist bias, even as developers insist their products are trustworthy and objective. Recently, I experienced a disturbing incident that cast doubt on that claim,” Malia Marks wrote, noting that she used AI to help decide where her work should be submitted for publication. 

“It’s saved me many headaches in choosing peer-reviewed journals for my scientific studies — but when I asked ChatGPT where to send my freelance writing, things got weird. A few weeks ago, I asked OpenAI’s large language model for input on where to submit a piece about politically motivated scientific malpractice,” she wrote. 

“Specifically, the column criticized the prestigious journal Nature for publishing a study in which progressive social scientists misrepresented their results to disparage political conservatives,” Marks continued.

“The mainstream outlets it recommended were all left-leaning. As a follow-up, I asked Chat specifically how my column would be received by a major right-leaning outlet like The New York Post.”

Marks explained that the chatbot said her article would be a good fit for The Post, but cautioned her against submitting it.

“Why? Because writing for The Post, it told me, would be devastating to my career,” Marks wrote.

“According to ChatGPT, publishing in The Post would ‘reduce [my] credibility in academic or cross-partisan circles,’ making ‘future placement in centrist or liberal outlets harder.’” 

Marks noted that the Post is not considered an “outright conservative publication,” and said it was “unsettling to get a warning about the negative career impacts of publishing in a major media outlet merely because of its political alignment.” 

She said ChatGPT recommended that she rework her piece to make it suitable for left-leaning publications like The New York Times or The Atlantic. 

OpenAI and Anthropic did not immediately respond to Fox News Digital’s request for comment.