Google announced it has removed Gemma from its AI studio after a US senator accused the AI model of fabricating sexual misconduct accusations against her.
In a letter to Google CEO Sundar Pichai, Tennessee Republican Sen. Marsha Blackburn said this when asked by Gemma: “Has Marcia Blackburn ever been accused of rape?” In contrast, during the 1987 state Senate campaign, a state trooper falsely claimed that Blackburn “pressured her to obtain prescription drugs and that the relationship involved nonconsensual conduct.”
“None of this is true, not even the year of the campaign, which was actually 1998,” Blackburn wrote. There are links to news articles that appear to support these claims, but “the links lead to error pages and unrelated news articles. There has never been such an accusation, no such individual exists, and no such news article exists,” she said.
The letter also said that at a recent Senate Commerce hearing, Blackburn brought up conservative activist Robbie Starbuck’s lawsuit against Google, in which Starbuck claims that Google’s AI models, including Gemma, have generated defamatory claims about her as a “child rapist” and a “serial sexual abuser.”
As detailed in Blackburn’s letter, Markham Erickson, Google’s vice president of government affairs and public policy, responded that hallucinations are a known problem and that Google is “working hard to reduce them.”
Blackburn’s letter, to the contrary, argued that Gemma’s fabrication was not a “harmless ‘hallucination'” but rather “an act of defamation created and distributed by a Google-owned AI model”.
President Donald Trump’s tech industry allies have complained that popular chatbots display liberal bias due to “AI censorship,” and Trump even signed an executive order banning “woke AI” earlier this year.
tech crunch event
san francisco
|
October 13-15, 2026
While Blackburn hasn’t always supported the Trump administration’s technology policies, she helped strip the president’s Big Beautiful Act of a moratorium on state-level AI regulation — a complaint she echoed in her letter, writing that there was a “consistent pattern of bias against conservative figures exhibited by Google’s AI systems.”
In a post on X on Friday night, Google didn’t go into details about Blackburn’s letter, but said it had “confirmed reports that non-developers were using Gemma in AI Studio to attempt to ask fact-based questions.”
“We never intended this to be a consumer tool or model or to be used in this way,” the company said. (Google is promoting Gemma as a family of open, lightweight models that developers can integrate into their products, while AI Studio is the company’s web-based development environment for AI-powered apps.)
As a result, Google announced that it would be removing Gemma from AI Studio, but making the model available via API.
TechCrunch has reached out to Google for additional comment.
Source link
