Gemma crossed the line
Google has pulled its Gemma AI model from its AI Studio platform after a US senator said the tool generated serious, and false, claims about her.
Google said Gemma was designed for developers, not the general public, and wasn’t meant to answer factual questions.
Because some people were using it in that way, the company removed access from AI Studio, though developers can still use Gemma through the API.
The complaint centres on Senator Marsha Blackburn, who said Gemma responded to a prompt by inventing allegations about her and even listed fake news links.
None of the details were accurate.
In brief:
Google removed Gemma from AI Studio after it generated false claims about a US senator.
The model was intended for developers only, not public factual queries.
The incident highlights ongoing accuracy issues (“hallucinations”) in AI systems.
Gemma wasn’t a Q&A girlie
The situation highlights an ongoing issue in generative AI: models can still produce incorrect information presented as fact.
Google said it remains focused on reducing these errors.
Blackburn has called for the model to be shut down until it can be better controlled.
Google really said “you’re holding it wrong” - MV