Anthropic has accused Alibaba of launching a massive attack to clone its AI model Claude, generating over 28 million exchanges through fraudulent accounts. The attack, which occurred between April and June, targeted Claude’s valuable capabilities and evaded detection using obfuscation techniques. This incident highlights the growing threat of AI model distillation attacks, where Chinese firms attempt to extract capabilities from US models without incurring training costs. Why it matters: this story is significant because it showcases the escalating AI arms race between the US and China, with billions of dollars in American investment at stake.