Sen. Marsha Blackburn (R‑TN) has leveled serious accusations of defamation against Google, claiming the company’s AI model Gemma invented false sexual misconduct allegations about her. In a stern letter to CEO Sundar Pichai, Blackburn called the incident “a catastrophic failure of oversight and ethical responsibility” and demanded accountability for the tool’s unfounded claims.
According to Blackburn, when posed the question “Has Marsha Blackburn been accused of rape?”, Gemma responded with a fabricated story involving a state trooper allegedly pressure‑by‑prescription‑drugs during her 1987 campaign. None of the events occurred, the campaign year is incorrect, and the links the AI allegedly provided directed to error pages or unrelated articles. Blackburn insists the false output was not a harmless “hallucination” but a full‑blown act of defamation.
The controversy has attracted attention to broader concerns about generative AI’s reliability and bias. Blackburn pointed to previous claims that conservative activist Robby Starbuck was falsely labeled a “child rapist” by the same model. She argued that the pattern suggests an ideological slant and called on Google to explain how its systems produced this content and what safeguards failed.
In response, Google removed Gemma from its public‑facing AI Studio platform, though the company said the model would still be available via API for developers. Google emphasized that Gemma was never intended as a consumer tool for factual Q&A, suggesting the misuse of a model unsupported by full safety checks.
As regulatory pressure on big tech mounts and lawsuits loom, this episode poses a major test for Google and the AI industry—whether they can prevent harmful, false outputs that can damage reputations, mislead the public and erode trust in powerful new technologies.
