AI · 2h ago
Anthropic accuses Alibaba of stealing Claude model via 25K fake accounts
Anthropic claims Alibaba used 25,000 fake accounts and 28.8 million exchanges to illicitly distill its Claude AI model from April to June 2026. The American startup traced the activity to operators affiliated with Alibaba and its Qwen team. The alleged theft involved systematically extracting model knowledge through repeated queries.
Meridian48 take
The scale of the alleged operation suggests a coordinated industrial espionage effort, raising questions about how Anthropic's safeguards failed to detect it earlier.
Read the full reporting
Anthropic claims that China's Alibaba used 25,000 fake accounts and 28.8 million exchanges to illicitly 'distill' its Claude model — violations occurred from April to June 2026 →
Tom's Hardware
model-distillationai-espionage