Author Wynton Hall used Google's AI chatbot Gemini to identify senators whose statements he said violate the platform's hate speech policies, finding that only Republican lawmakers were flagged by the system.

Hall used Gemini Pro's "deep research" function and presented his findings to Fox News Digital. The AI flagged several Republican senators but identified no Democrats as violating its hate speech guidelines.

Among those flagged was Sen. Marsha Blackburn of Tennessee, whom Gemini cited for characterizing 'transgender identity as a harmful cultural influence' and using 'woke' as what the AI characterized as a derogatory slur against protected groups. Sen. Tom Cotton of Arkansas was also flagged for cosponsoring legislation 'to exclude transgender students from sports.'

Google did not immediately respond to Fox News Digital's request for comment on the findings.

What the Right Is Saying

Hall argues that his findings demonstrate a systemic bias in AI systems that reflects the political leanings of their creators. His new book, 'Code Red: The Left, The Right, China and the Race to Control AI,' argues that AI tools marketed as neutral are shaped by the ideological assumptions of their developers.

'AI's Silicon Valley architects lean left politically, and their lopsided political donations to Democrats underscore their ideological aims,' Hall told Fox News Digital.

Hall writes that AI has put Big Tech's consolidating control 'on steroids,' arguing that users often trust AI outputs too much without recognizing potential bias. He points to what he describes as a 'closed loop' where AI systems are trained on content from legacy media outlets that he says largely exclude conservative perspectives.

Hall argues that conservatives must respond by demanding transparency in AI training data and algorithmic accountability. He cites PayPal co-founder Peter Thiel's characterization of Silicon Valley as 'a one-party state' as evidence of the ideological environment shaping AI development.

What the Left Is Saying

Progressive critics and tech policy experts have challenged Hall's characterization of AI bias as systemic. They argue that focusing on individual outputs misses the broader picture of how AI systems are trained and evaluated.

Democrats and progressive tech watchdogs have noted that Silicon Valley companies, including Google, have donated to both parties and have increasingly engaged with conservative lawmakers on regulatory issues. They also point out that tech companies have made significant donations to presidential inaugurations, including Trump's 2024 inauguration.

Some progressive analysts argue that AI systems reflect the diversity of perspectives in their training data, which includes academic consensus on issues like civil rights and historical context. They contend that flagging content that discriminates against protected groups is not partisan but rather a factual application of anti-discrimination standards.

What the Numbers Show

Hall claims that 85% of political donations from employees at Apple, Meta, Amazon and Google go to Democrats, based on his research cited in the book. This figure aligns with some independent analyses of tech employee political giving.

The source article notes that major tech companies made $1 million donations to Trump's 2024 inauguration, following customary practice. Hall argues these gestures did little to hide where Silicon Valley's loyalties had long been positioned.

The article also references specific examples of rhetoric from some Democrats, including a 2023 statement by Rep. Dan Goldman of New York who said then-candidate Trump was 'destructive to our democracy' and needed to be 'eliminated,' for which he later apologized. In 2024, Texas Democratic House candidate Jolanda Jones made a throat-slashing gesture on CNN while discussing political conflict.

These examples were presented in the context of questioning why Gemini flagged certain Republican statements but not these Democratic examples, though it is unclear whether Hall specifically prompted Gemini to evaluate these statements.

The Bottom Line

The finding from a single researcher using one AI tool's deep research function raises questions about content moderation consistency across platforms. Google has not responded to the specific allegations in this case.

The broader debate over AI bias reflects ongoing tensions between tech companies, policymakers and the public over content moderation standards. Both progressive and conservative critics have raised concerns about algorithmic decision-making in ways that reflect their respective priorities.

As AI systems become more integrated into information discovery, questions about training data transparency and moderation consistency will likely remain a point of political contention. Users should evaluate AI outputs critically and consider multiple sources when forming opinions on contested topics.