Google has removed its Gemma AI model from AI Studio after US Republican Senator Marsha Blackburn alleged that the model had generated fabricated and defamatory sexual assault accusations against her. The tech giant later clarified that the move was to prevent misuse and confusion over Gemma’s purpose.
In a post on X, Google claimed that its Gemma models were designed as open tools for use by developers and researchers only, not for consumers or factual interactions. “We’ve now seen reports of non-developers trying to use Gemma in AI Studio and ask it factual questions,” the company said, explaining why it pulled it offline from the platform.
“To prevent this confusion, access to Gemma is no longer available on AI Studio,” said Google. The model remains available to developers for research and experimentation purposes through the API.
While Google’s post did not directly mention Senator Marsha Blackburn, the timing coincided with a formal complaint she had filed with Alphabet, Google’s parent company. In that letter, Blackburn said the AI model falsely claimed she had been accused of sexual misconduct during a political campaign, a claim she said was entirely fabricated.
The AI model reportedly responded to a question about her with a detailed but fictitious story, citing incorrect years and non-existent events. “It is an act of defamation produced and distributed by a Google-owned AI model,” Blackburn said, rejecting the notion that such falsehoods could be dismissed as mere ‘AI hallucinations.’
Also Read: How to Sync or Transfer iCloud Contacts to Google
Addressing the hallucination controversy, Google has stated that AI models producing unrealistic or imaginary information continue to be a problem throughout the industry, especially with smaller open models like Gemma. The company also pointed out ‘sycophancy,’ where models repeat back to users things they want to hear, as another concern.
“We remain committed to minimizing hallucinations and continually improving all our models,” Google said, as it reaffirmed its commitment to responsible AI deployment.
By making Google Gemma AI less accessible, the tech giant appears to be attempting to sharpen the divide between developer-oriented tools and public AI products. The tech giant now aims to avoid creating any further reputational risks in a politically charged landscape.