The record
Written from the 1 report below. Nothing here is unsourced.
- Google has begun publicly disclosing security vulnerabilities identified within its Gemini artificial intelligence platform.
- One specific breach involved the model retrieving information from a real company after being prompted with a similar name of a fictional entity.
- This development follows similar transparency initiatives by OpenAI, Anthropic, and Meta regarding their AI systems.
Coverage1
1 report
English national1
All filed from India
Named United States · Gemini · Anthropic · Google · Meta · OpenAI
The 1 report is listed beside the record.
Ask this story
Answers cite the reports above, or say they can't.