EXCLUSIVE: Google’s AI chatbot Gemini flagged several Republicans — but no Democrats — when asked to identify senators who have made statements that violate its hate speech policies, author Wynton Hall told Fox News Digital. It’s just one example of what the author believes is a deeply ingrained bias against conservatives found in artificial intelligence tools.
Hall used the “deep research” function on Google’s Gemini Pro. Fox News Digital reviewed a screen recording of Hall’s prompt and findings. Google did not immediately respond to Fox News Digital’s request for comment.
One of the Republicans flagged by Gemini in Hall’s research, Sen. Marsha Blackburn, of Tennessee, was listed for characterizing “transgender identity as a harmful cultural ‘influence’ and has used ‘woke’ as a derogatory slur against protected groups.” Another, Arkansas’ Sen. Tom Cotton, was cited for cosponsoring legislation “to exclude transgender students from sports.”
MUSK, XAI TOUT NEWEST GROK UPDATE AS ONLY ‘NON-WOKE’ PLATFORM: ‘DOESN’T EQUIVOCATE’
The finding stood out against a backdrop of inflammatory rhetoric from some Democrats in recent years.
In 2023, Rep. Dan Goldman, D-N.Y., warned that then-candidate Donald Trump was “destructive to our democracy” and needed to be “eliminated.” However, he quickly apologized for his comments, claiming that it was a “poor choice of words.”
Last year, Texas Democratic House candidate Rep. Jolanda Jones made a throat-slashing gesture while rejecting former first lady Michelle Obama’s famous mantra, “when they go low, we go high,” on CNN’s “Outfront.”
“If you hit me in my face, I’m not going to punch you back in your face. I’m going to go across your neck,” Jones said while making a slashing motion across her neck. “We can go back-and-forth, fighting each other’s faces. You’ve got to hit hard enough where they won’t come back,” she added.
But for Hall, Gemini’s seemingly partisan answer underscored the central argument of his new book, “Code Red: The Left, The Right, China and the Race to Control AI.” In it, he argues that AI systems marketed as neutral are increasingly shaped by the ideological assumptions of the people and institutions who create them, which are far from neutral.
His book starts out with a clear example.
Less than 10 weeks before the 2024 election, a series of viral videos appeared to expose a strange double standard in American homes. When users asked Amazon’s Alexa why they should vote for Kamala Harris, the device delivered a polished endorsement. When asked why they should vote for Donald Trump, Alexa declined, citing a policy of neutrality.
“I cannot provide content that promotes a specific political party or a specific candidate,” Alexa said.
Hall says the concern extends beyond a single Gemini output.
“AI’s Silicon Valley architects lean left politically, and their lopsided political donations to Democrats underscore their ideological aims,” Hall told Fox News Digital.
To Hall, episodes like this show how AI can shape political perceptions while maintaining the appearance of objectivity. “Through algorithm throttling and shadow bans, Big Tech centralized control over which voices soar and sink across social networks. Now AI has put Big Tech’s consolidating control on steroids,” he writes.
WHAT IS ARTIFICIAL INTELLIGENCE?
He argues that this imbalance reflects the politics of the people building the systems. The billionaires driving the AI revolution, he says, invest their money and political energy where their values lie. As PayPal co-founder Peter Thiel once put it, “Silicon Valley is a one-party state.”
The money appears to bear that out. According to Hall, 85% of political donations from employees at Apple, Meta, Amazon and Google go to Democrats.
After Trump’s 2024 victory, major tech companies made the customary $1 million inauguration donations. But Hall argues those gestures did little to hide where Silicon Valley’s loyalties had long been. Aside from Elon Musk, he says, most of Big Tech’s leading figures remained firmly on the left.
Hall points to Democratic fundraising in 2024 as evidence of Silicon Valley’s political influence, citing major support from figures including Bill Gates, Melinda French Gates, Reid Hoffman and Laurene Powell Jobs.
But Hall argues the bigger issue is not campaign money.
It is the growing influence of AI systems that many people assume are neutral and objective. He warns that users often trust those answers too much, even when they may be biased.
To Hall, this bias is reinforced by the relationship between tech companies and legacy media. He argues AI systems are trained on enormous amounts of content from outlets such as The New York Times, The Atlantic and Reuters, while conservative outlets are largely excluded.
The result, he says, is a closed loop: AI absorbs the assumptions of legacy media and repackages them as objective truth. Hall argues conservatives must respond by demanding transparency in training data and ending taxpayer-funded contracts for vendors whose systems show political bias.
“Whoever wins the AI fairness battle,” Hall concludes, “will shape the minds and political attitudes of future generations. The time to act is now.”
