Alibaba accused of large‑scale AI theft from Anthropic

Anthropic, a U.S. AI startup, alleges that China’s Alibaba and its affiliated AI research unit launched the largest known "distillation attack" on its Claude model. Between April 22 and June 5, the group allegedly used about 25,000 fraudulent accounts to conduct over 28.8 million interactions with Claude, extracting data to train a rival system.

Anthropic reported the breach in a letter to U.S. Senators Tim Scott and Elizabeth Warren, describing the activity as a serious intellectual‑property violation and urging government action. The case highlights rising tensions over AI technology transfer and the need for stronger safeguards against large‑scale model theft.