shared a link post in group #🇨🇳 ChinA.I. 🤖🧠🦾🤖
In a blog post on Monday, the AI giant accuses three high-profile Chinese companies — DeepSeek, Moonshot AI, and MiniMax — of using Claude’s system and information to train and refine their own models, via a process called “distillation.” While there are perfectly white hat reasons to distill large AI models — producing smaller and more fast/efficient alternatives, for example — according to Anthropic, their open-source Chinese rivals are using the tactic to re-purpose their work into new products, at lower costs. According to Anthropic’s data, these three Chinese firms created more than 24,000 “fraudulent” Claude accounts, while engaging in hundreds of thousands or even millions of interactions each. Anthropic’s not the only US AI company making this case; OpenAI sent a memo to lawmakers just weeks ago making similar accusations against DeepSeek specifically.
#🇨🇳 ChinA.I. 🤖🧠🦾🤖
https://www.wsj.com/tech/..
