Google has removed its Gemma artificial intelligence model from AI Studio. This action follows allegations made by U.S. Senator Marsha Blackburn after the AI incorrectly asserted serious claims of sexual misconduct against her.
Allegations Against Gemma AI Model
Senator Blackburn raised her concerns in a letter directed to Google CEO Sundar Pichai. She highlighted details regarding the AI’s false response to a question about her, which claimed that she had been accused of nonconsensual acts during a 1987 state senate campaign.
Key Points of the Allegation
- The AI model stated Blackburn pressured a state trooper for prescription drugs.
- It falsely referred to a 1987 campaign, while the actual year was 1998.
- Gemma generated fake links allegedly supporting its claims, which led to error pages.
Senator Blackburn vehemently denied these accusations, emphasizing that they are entirely false and defamatory. She insisted that such inaccuracies represent a serious failure of oversight and ethical responsibility on Google’s part.
Response from Google
During a Senate Commerce hearing titled “Shut Your App,” Blackburn criticized Google’s handling of AI outputs that fabricate serious claims against conservative figures. Google’s Vice President for Government Affairs, Markham Erickson, acknowledged that “hallucinations” occur in large language models and stated the company is working to mitigate these issues.
Concerns About Bias in AI Models
Blackburn pointed out a troubling pattern where conservative individuals are disproportionately targeted by defamatory content generated by Google’s AI systems. The senator requested a substantive written response from Google addressing the following points by November 6, 2025:
- An explanation of how the AI generated false accusations against her.
- Steps taken to identify and eliminate ideological bias during model training.
- Information about internal testing and safeguards to prevent AIgenerated libel.
- Concrete measures to remove defamatory content from the Gemma model.
Blackburn concluded by emphasizing the necessity for AI systems that uphold accuracy, fairness, and transparency. She insisted that tools producing baseless allegations against public figures must be effectively controlled before widespread use.