Anthropic Discloses Industrial-Scale Distillation Attacks: DeepSeek, Moonshot AI, and MiniMax Created 24K Fraudulent Accounts Generating 16M Exchanges

X/Twittervia @AnthropicAI·high signal

Anthropic officially named three Chinese AI labs — DeepSeek, Moonshot AI, and MiniMax — as having orchestrated coordinated model distillation attacks using 24,000 fraudulent accounts that generated 16 million exchanges to extract Claude's outputs at scale. The announcement drew 33.5 million views and 55,000 likes, making it one of the most-seen AI industry disclosures in recent memory. This is the first time Anthropic has publicly named specific competitors and quantified the scope of IP extraction, transforming a long-rumored industry practice into a named, documented attack campaign.