Additionally, Google has withdrawn its AI model, Gemma, from the AI Studio. This latest action comes on the heels of Senator Marsha Blackburn’s charge that the entire system is defamatory. The conflict arose when Blackburn highlighted that the model produced false information about her during her 1998 state senate campaign, which she characterized as “an act of defamation produced and distributed by a Google-owned AI model.”
Blackburn, along with New York City Mayor Eric Adams and others, took to writing letters to Google to draw attention to their misstep. Second, Gemma misidentified the campaign year as 1987, when it was actually 1998. She contended that the AI’s total fabrications amounted to “not a harmless ‘hallucination.’” Rather, they constituted a serious defamation that would injure her professional reputation.
Markham Erickson, Google’s Vice President for Government Affairs and Public Policy responded on Google’s Blog and took on some of Blackburn’s criticisms. He reaffirmed that the company is working to address the problems associated with AI-created hallucinations. He acknowledged that non-developers have already been testing Gemma in the AI Studio for fact-based queries. That wasn’t the model’s original use case.
“Has Marsha Blackburn been accused of rape?” – Excerpt from Blackburn’s letter
Blackburn’s criticism of Google’s AI system not her first time getting involved in tech policy. Although she has not always supported the Trump administration’s tech initiatives, she played a pivotal role in stripping a moratorium on state-level AI regulation from Trump’s “Big Beautiful Bill.” Earlier this year, she warned about the growing threat of “AI censorship.” She launched attacks against what she termed “woke AI”—part of a deepening general mistrust of tech firms.
This incident further highlights the critical question of how AI systems can be held accountable and the dangers of biases within AI algorithms. Blackburn described a “consistent pattern of bias against conservative figures demonstrated by Google’s AI systems,” suggesting that such biases could influence public perception and political discourse.
Google has repeatedly insisted that it never meant for Gemma to be a tool for consumers. The company is committed to addressing the threats posed by generative AI technologies. It doesn’t just passively hope that its models give good results.
“We never intended this to be a consumer tool or model, or to be used this way” – Google
The repercussions of this scenario are profound, extending well beyond the borders of Blackburn. More importantly, it highlights mounting opposition among lawmakers and tech leaders to the perceived misuses of artificial intelligence. In the developing AI regulation debate, both Democrats and Republicans will be judged on how they handle these difficult and nuanced topics.