NASHVILLE, Tenn. — U.S. Senator Marsha Blackburn is demanding accountability from Google after its artificial intelligence model, Gemma, generated fabricated criminal allegations against her — an incident that led the tech giant to pull the model from its AI Studio platform late last week.
In a letter to Google CEO Sundar Pichai, Blackburn said the company’s large language model produced false claims that she had been accused of rape during a 1987 campaign for the Tennessee Senate, citing nonexistent sources and fake links to fabricated news articles. The senator called the episode “an act of defamation produced and distributed by a Google-owned AI model.”
“Whether intentional or the result of ideologically biased training data, the effect is the same,” Blackburn wrote. “Google’s AI models are shaping dangerous political narratives by spreading falsehoods about conservatives and eroding public trust.”
The letter followed a Senate Commerce Committee hearing in which Blackburn questioned Google’s vice president for government affairs, Markham Erickson, about the company’s inability to prevent its AI systems from generating malicious content about public figures. Erickson acknowledged that “hallucinations” are a known issue in large language models and said Google was “working hard to mitigate them.”
Blackburn rejected that explanation, saying such failures go beyond technical error and represent a “catastrophic failure of oversight and ethical responsibility.” She urged the company to suspend the system entirely until it could ensure accuracy and neutrality, declaring, “Shut it down until you can control it.”
According to reporting from TechCrunch, Google confirmed Friday that it had removed Gemma from its AI Studio interface after reports surfaced of non-developers using the model to ask factual questions. In a statement on X, Google said Gemma “was never intended to be a consumer tool or model” and that its continued availability will be limited to developers through API access.
The incident comes amid growing political scrutiny over generative AI and its potential to spread misinformation or defamatory content. Blackburn, who has frequently accused major tech companies of bias against conservatives, said the episode demonstrates the dangers of releasing advanced AI systems without sufficient safeguards.
“Google has a long history of targeting, censoring, and smearing conservatives,” Blackburn said in a follow-up statement after Gemma’s removal. “It is absolutely unacceptable that Google did not prevent its large language model from pulling heinous criminal allegations out of thin air and presenting them as fact to unknowing users.”
The controversy also references conservative activist Robby Starbuck’s lawsuit against Google, alleging similar false claims generated by Gemma and other AI models. Blackburn cited the Starbuck case during last week’s Commerce hearing as evidence of a broader pattern of bias within Google’s AI ecosystem.
Google has not directly addressed Blackburn’s accusations of political bias but said it continues to refine its AI systems to minimize factual inaccuracies. The company has not yet released a public response to Blackburn’s letter.
The Tennessee senator says she intends to press for formal oversight measures to ensure AI developers are held accountable when their systems produce false or defamatory statements about real individuals.
Discover more from TNPOLITICO
Subscribe to get the latest posts sent to your email.
