Claude distillation claims raise cost and security questions
Anthropic accused Alibaba and its Qwen AI lab of trying to copy Claude’s abilities for competing . The claim says operators used nearly 25,000 fake accounts between April and June 2026. Those accounts created 28.8 million interactions with Claude.
The alleged goal was to gather enough Claude output to train other models, a described as a . Alibaba has not publicly answered the accusation, and the claims have not been independently confirmed.
Key points
- Anthropic accused Alibaba and its Qwen AI lab of trying to extract Claude’s abilities.
- The alleged activity ran from April to June 2026.
- The claim involves nearly 25,000 fake accounts and 28.8 million Claude interactions.
- The alleged aim was to train competing using Claude output.
- Alibaba has not publicly responded, and the claim is not independently verified.