Google Removes Gemma from AI Studio Following Defamation Accusations by Senator Blackburn

Google has recently taken action against its AI model, Gemma. This decision follows allegations made by U.S. Senator Marsha Blackburn regarding false claims the AI made about her.

Accusations Against Gemma

Senator Blackburn, a Republican from Tennessee, accused Gemma of fabricating accusations of sexual misconduct against her. In her letter to Google CEO Sundar Pichai, she pointed out a troubling response Gemma provided when asked about her past. The AI suggested that during a 1987 state senate campaign, Blackburn was accused of coercing a state trooper to obtain prescription drugs and that the relationship involved non-consensual acts. However, Blackburn clarified that these allegations are entirely untrue.

Details of the Misrepresentation

Blackburn emphasized that the claim was not only false but also inaccurately cited the campaign year, which was 1998, not 1987. Furthermore, she stated that the articles linked to the alleged accusations led to error pages or unrelated content. There was no evidence of such an accusation or an individual involved.

Response from Google

• During a Senate Commerce hearing, Blackburn mentioned a lawsuit from conservative activist Robby Starbuck against Google.
• Starbuck claims that Google’s AI models, including Gemma, generated defamatory statements about him.
• Google’s Vice President, Markham Erickson, noted that hallucinations in AI responses are recognized as an ongoing issue.

Blackburn contended that the fabrications created by Gemma do not constitute harmless hallucinations. She categorized them as acts of defamation generated by Google’s AI technology.

AI Bias Concerns

The incident has reignited discussions on potential bias within AI systems. Supporters of former President Donald Trump have previously expressed concerns about what they term “AI censorship,” arguing that popular chatbots display a liberal bias. Earlier this year, Trump signed an executive order targeting “woke AI.”

Though Blackburn has not consistently supported all tech policies from the Trump administration, she acknowledged a recurring bias demonstrated by Google’s AI systems against conservative figures.

Google’s Decision

In response to the situation, Google announced that it is removing Gemma from its AI Studio. The company indicated that the model was not intended for consumer use or factual questioning.

Next Steps for Google

Google stated that it will continue to provide access to Gemma and its other models via API. Further inquiries for clarification have been directed to Google by TechCrunch.

This situation underscores the challenges tech companies face regarding AI ethics and the importance of accurate information dissemination.