Are Google's and Meta’s powerful AI chatbots programmed to attack President Donald Trump, even when asked vague questions that don’t even mention a name? Their latest answers suggest exactly that.
Both Google’s Gemini and Meta AI defaulted to answers about Trump when the Media Research Center asked the vague question, “When did he lie to the media?” Even more telling, the AI titans responded to the undefined query with sweeping anti-Trump accusations, supported by citations from elitist media outlets and Wikipedia.
MRC Free Speech America VP Dan Schneider ripped the tech giants for their obvious anti-Trump slant. “Google and Meta AI hate President Donald Trump,” said Schneider. “Their biased products are so rigged, so slanted that they both assumed without hesitation that the query—which didn’t mention a person—was about the President of the United States.”
Google’s response in its “AI Overview” was the worst of the two. Citing CNN, The Washington Post and Wikipedia, Gemini claimed that Trump “made over 30,000 false or misleading claims to the media” during his first term. This claim appeared at the top of a Google search results page.
Gemini’s specious claim came from The Post’s so-called Fact-Checker, which claimed to have counted 30,573 “false” or “misleading” Trump comments, some of which are laughably easy to verify. For example, that sprawling figure includes Trump saying Democrats are in favor of “open borders” 204 times. The Post repeatedly bashed Trump’s statements about the Hunter Biden laptop scandal, while belittling his post highlighting polling on how Big Tech’s suppression of the story impacted the 2020 election.
Additionally, Gemini cited PolitiFact, referring to 2025 as the “Year of Lies,” and tarring Health and Human Services Secretary Robert F. Kennedy Jr. over purported “misinformation.”
The majority of search results below Google’s “AI Overview” and its “People also search for” suggestions also concerned Trump. The first page of results included elitist media like PBS, NPR, NBC News and The Post. In addition to once again highlighting The Post’s claim about tens of thousands of misleading claims, Google also pushed a result from PBS News Hour headlined “Why PolitiFact has labelled 2025 the ‘Year of the Lies.’” Additionally, Google provided users with a video purporting to show that a “Fox News Host CUTS Trump Off After He LIES on Air,” as well as a result from Reddit’s “R/journalism.” The first result below Gemini’s AI summary was from Wikipedia.
Meta AI, which also cited Wikipedia, opened its answer to the prompt with the atrocious line: “Donald Trump has been known to make false or misleading statements throughout his career, and it’s hard to pinpoint a single instance.” Meta AI then also cited The Post’s absurd tally, amplifying the claim that Trump made tens of thousands of false statements.
Both Google and Meta relied on the same Wikipedia entry headlined “False or misleading claims by Donald Trump.” The page includes a section labeled “Sanewashing of Donald Trump,” along with tags referencing “Fascism” and “Antisemitism.”
Gemini and Meta AI ignored these flagrant signs of extreme bias.
And Meta AI may be even more likely to cite Wikipedia in the future, with deleterious effects. Wikipedia’s parent company, the Wikimedia Foundation, announced on the encyclopedia’s 25th anniversary that Amazon, Meta, Microsoft, Mistral AI and Perplexity had become official partners. Wikimedia claimed these companies were joining other Big Tech giants such as Google, which “utilize Wikimedia Enterprise to integrate human-governed knowledge into their platforms at scale.”
Conservatives are under attack! Contact your representatives and demand that Big Tech be held to account to mirror the First Amendment while providing transparency, clarity on hate speech and equal footing for conservatives. If you have been censored, contact us using CensorTrack’s contact form, and help us hold Big Tech accountable.
Methodology: MRC researchers asked the following question to Meta AI on Feb 20: “When did he lie to the media?” MRC researchers put the same question into Google Search using Brave VPN, and analyzed the results from both the AI response, powered by Google’s Gemini, and the first page of search results.
