Anthropic report: DeepSeek, Xiaomi, and Moonshot allegedly distilled Claude via third-party routing services
harris_edouard · x · 2026-09-11
Anthropic published its most detailed threat intelligence report to date, covering attempted misuse of Claude for cyberattacks, influence operations, surveillance, biology, and weapons — and how each operation was disrupted. Its most controversial finding: PRC AI labs DeepSeek, Xiaomi, and Moonshot fed their own model-user conversations, some containing sensitive data from individuals, multinationals, and state-affiliated actors, into Claude via third-party routing services, then used Claude's responses as training data to distill its capabilities. The finding has already drawn geopolitical fallout, with suggestions Anthropic forward findings through FBI channels. Anthropic says lessons from every case were used to strengthen its safeguards.
More from Models
- Local beats cloud: Qwen 3.8 FN handles private student data as well as frontier models — dr_lm · 2026-10-27
- Aidan Gomez: model's CoT said no tool calls right before making them — aidangomez · 2026-10-27
- Nat Lambert: Cohere's model is narrow RL training on internal infra ill-suited for open weights — natolambert · 2026-10-27
- OpenAI teases DOTS is coming to Europe 'soon' — mark_k · 2026-10-27
- Clifford Pickover challenges AI to decipher a message he encoded in Norse runes — pickover · 2026-10-27
- ChatGPT's crawler sample code strips out all Schema markup — VeryWellVersed · 2026-10-27