OpenAI has identified a campaign involving the unauthorized extraction of protected AI reasoning, linking the core activity to individuals associated with the Chinese developer Moonshot AI.
OpenAI has disrupted a campaign linked to Moonshot AI that attempted to extract and copy the internal reasoning processes of its AI models.
OpenAI and Anthropic have accused Chinese lab Moonshot AI of using fraudulent accounts to extract protected reasoning data from their models for distillation purposes.
U.S. AI labs are pushing for government intervention against distillation attacks as they prepare for high-level talks with China regarding the future of artificial intelligence development.
Anthropic reports that Chinese AI firms, including Alibaba and Moonshot AI, have launched aggressive distillation campaigns to harvest capabilities from Claude models.
