Anthropic and other companies allege Chinese firms used thousands of fake accounts to replicate advanced US AI models at low cost.
Anthropic, a $1 trillion-valued AI firm, has formally accused Chinese companies of stealing intellectual property through tens of thousands of unauthorized accounts. The allegations, detailed in a letter to US Senators, claim Chinese firms used a process called ‘distillation’ to replicate advanced AI models nearly as powerful as US counterparts at minimal expense.
The method involves creating fake accounts to access and aggregate proprietary AI technologies. Alibaba and other major Chinese tech firms are implicated, raising concerns about both economic competition and national security risks. Experts warn the stolen models could also be used to extract sensitive data from US companies.
The accusations highlight escalating tensions over AI dominance, compounded by China’s advancements in semiconductor technology to rival Nvidia’s chips. The revelations may prompt stricter US regulatory measures on AI exports and collaborations.