Anthropic claims that Alibaba is conducting the largest distillation campaign aimed at Claude.
TL;DR: Anthropic has accused Alibaba's Qwen lab of conducting the largest distillation campaign against a U.S. AI company, claiming that nearly 25,000 fake accounts were utilized to carry out approximately 29 million exchanges with Claude from April to June. This marks the first time Anthropic has identified a major Chinese tech company as engaging in a distillation attack, previously pointing to smaller Chinese AI startups. Distillation involves using structured queries to gather responses from a leading AI model and training a rival system. The White House has expressed concerns about this technique as a national security issue. In response to the allegations, Alibaba's stock fell over three percent. Separately, officials in Washington are exploring sanctions against any Chinese firms deemed to be improperly accessing U.S. AI outputs, while Anthropic is facing its own regulatory challenges with the U.S. government. The overarching issue could influence the future regulatory landscape for U.S. AI firms and the competitive dynamics between them and Chinese laboratories.
Other articles
Anthropic claims that Alibaba is conducting the largest distillation campaign aimed at Claude.
Anthropic informed US senators that Alibaba's Qwen lab utilized 25,000 fraudulent accounts to conduct nearly 29 million interactions with Claude from April to June.
