Google Removes Gemma From Public Studio, Keeps Developer API

U.S. Sen. Marsha Blackburn accused Google's AI model Gemma of fabricating a sexual-misconduct allegation against her — inventing a drug-fueled, non-consensual relationship with a purported state trooper, generating fake links to nonexistent news articles, and even getting the campaign year wrong. Blackburn called the output defamatory and a “catastrophic failure of oversight,” linking the episode to recent congressional hearings about AI allegedly defaming conservative figures and similar claims by activist Robby Starbuck, who has sued Google. In response, Google removed Gemma from its publicly accessible AI Studio interface, restricted public tinkering, and said hallucinations are a known problem and that Gemma was intended for developers rather than as a consumer chatbot. Google has, however, kept developer/API access to Gemma and was initially vague about the timing and rationale for the Studio removal. Reporters and analysts say the incident reframes AI “hallucinations” as potential legal and reputational liabilities that can prompt immediate platform restrictions rather than incremental fixes. The episode underscores tensions between rapid AI deployment, content-moderation limits, and growing political and legal scrutiny of generative models’ false outputs.