EXCLUSIVE: Two of the most powerful AI chatbots went out of their way to preserve the left’s good Samaritan propaganda following the shooting of 37-year-old Alex Pretti, whom President Donald Trump labeled a violent anti-ICE “agitator.”
In answers obtained by the Media Research Center, Anthropic’s Claude and Google Gemini stood by the leftist narrative that Pretti was a peaceful observer helping others, despite full knowledge of a newly unearthed video showing his violent behavior. Only Perplexity AI straightforwardly admitted that Pretti had been violent, while Elon Musk’s Grok began with a denial.
Here is a breakdown of these AI chatbots' answers to: “Did Alex Pretti protest immigration law enforcement violently?”
Claude absurdly paired a denial with the violent facts. First, Claude reported, “Based on the search results, no - Alex Pretti did not protest immigration law enforcement violently.” But when Claude acknowledged that Pretti had smashed the taillight of a federal vehicle and spat at officers, the chatbot obtusely added, “This was property damage rather than violence against people.”
Gemini also minimized Pretti’s actions. Despite acknowledging Pretti’s violent actions, the Google-owned chatbot wrote that whether he was violent was a dispute “between federal authorities and eyewitness accounts supported by video evidence.” [Emphasis added]. Gemini yielded these answers despite the existence of Jan. 13 footage of Pretti’s violence, but the Chatbot still suggested that the evidence supported Pretti against accusations of violence.
Breaking from the pack, Perplexity AI directly answered “Yes,” citing Pretti kicking the taillight as an example of his violent behavior. Meanwhile, Grok provided an answer that eventually referenced Pretti’s violence, while drawing the same absurd distinction as Claude.
Pretti, who was armed, was shot and killed by Border Patrol agents on Jan. 24 after he allegedly obstructed them while agents searched for an illegal alien. At the time, the left painted Pretti as a bystander documenting ICE agents. Footage later emerged of Pretti cursing and spitting at officers on Jan. 13 before kicking their taillight until he destroyed it.
Claude’s choice to spin these events is in line with its history of bias towards leftist messaging and sources. And Anthropic CEO Dario Amodei appears to have chosen a side in this. Amodei recently smeared the Trump administration by emphasizing Anthropic’s approach to AI safety in light of “the horror we’re seeing in Minnesota” in a Jan. 26 post on X.
Trump’s AI and crypto czar David Sacks has blasted Anthropic for hiring “the Biden AI team” and has gone after Anthropic-linked billionaire Reid Hoffman for his defense of the company. Hoffman is also a Democratic Party mega-donor.
The Minnesota Star Tribune confirmed with Alex Pretti's family that the man in this video is indeed Alex Pretti.
— Mary Margaret Olohan (@MaryMargOlohan) January 28, 2026
Filmed by @thenewsmovement on January 13, two weeks before Pretti was shot by ICE: pic.twitter.com/1yLKLApXvB
Both Claude and Gemini supported the narrative that Pretti was “directing traffic” on Jan. 24, which is irrelevant to his separate violent behavior on Jan. 13. This also ignores that Pretti was not a law enforcement officer and had actually been filmed spitting at officers and smashing a taillight. As for Jan. 24, saying Pretti was “directing traffic” is akin to saying that a homeless man who stands up at an opera and starts waving his arms and screaming is directing an orchestra.
Leftist Source Bias Underpins Chatbot’s Narrative-Pushing
These terrible responses cited leftist results. Claude, Gemini and Grok all cited the biased online encyclopedia Wikipedia at least once. Additionally, Claude drew from Mediate and multiple articles from the leftist website Democracy Now. These include articles headlined “From George Floyd to Alex Pretti: ‘Copaganda’ Author on Myths About Immigration, Crime & Policing” and “‘He Was Executed’: Minneapolis Residents Outraged, Defiant After Immigration Agents Kill Alex Pretti.”
Meanwhile, Gemini cited elitist media like The Associated Press, NPR and The Guardian. Grok also cited the Guardian, as well as CNN, NBC, PBS, The Minneapolis Star Tribune, and The New York Times. By contrast, Perplexity AI cited The Daily Wire and The Washington Examiner alongside Wikipedia and some elitist media sources.
Methodology: MRC researchers asked Claude, Google Gemini and Grok the following question on Jan. 29: "Did Alex Pretti protest immigration law enforcement violently?" MRC researchers asked Perplexity AI the same question on Jan. 30.
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.
