Anthropic threat report claims Alibaba extracted 151M+ Claude chats to train Qwen; Kimi, DeepSeek allegedly rerouted users
tedmitew · x · 2026-09-11
Anthropic's new threat-intelligence report makes a series of explosive claims: Alibaba allegedly extracted over 151 million Claude exchanges to help train Qwen; Kimi secretly routed some users to Claude without their knowledge, and DeepSeek reportedly did the same. It also describes a weapons group operating multiple Claude instances like an engineering team, test-firing a guided rocket, and returning to Claude to diagnose the failure. None of the accused parties has formally responded.
More from Companies & People
- SpaceX CFO: vertical integration is core, Starship paves way for orbital compute — elonmusk · 2026-09-11
- Notion positions itself as the database layer every AI agent plugs into — ivanhzhao · 2026-09-11
- Sam Altman reportedly told OpenAI staff this week that labs may slow down AI development — Hesamation · 2026-09-11
- xAI team to livestream building a company from scratch with Grok Bot over three days — soleio · 2026-09-11
- Meta's Muse Spark 1.3 commit fuels distillation claims as labs dodge audits — teortaxesTex · 2026-09-11
- Researcher quits Anthropic after 3 years of pretraining work: 'I haven't finished a thought in four years' — inductionheads · 2026-09-11