US says Chinese firms extracted billions of tokens from frontier AI models
By Bill Toulas
AI Summary
U.S. cybersecurity and intelligence agencies formally accused six Chinese AI companies of conducting industrial-scale 'distillation attacks' against American frontier AI models including those from OpenAI, Anthropic, Google Gemini, and SpaceX's Grok, extracting billions of tokens since at least late 2024. The technique involves systematically querying frontier models to capture outputs and reasoning processes, which are then used to train competing models at a fraction of the development cost. The U.S. government response includes urging AI firms to secretly identify and downgrade Chinese users to less-capable model versions, raising significant policy and operational security implications for AI providers.
Relevance score: 77.0/100
Expert cybersecurity solutions to safeguard your organization from evolving threats.
Get Protected →