A US artificial intelligence giant has accused Chinese tech firm Alibaba of stealing its AI capabilities through a massive extraction campaign, with nearly 29 million exchanges made using thousands of fraudulent accounts.
The San Francisco-based company, Anthropic, sent a letter to US Senators Tim Scott and Elizabeth Warren, detailing the alleged theft of its Claude AI model’s capabilities by operators linked to Alibaba.
Artificial Intelligence Theft
According to Anthropic, the campaign was carried out through “distillation attacks”, which extracted answers from a stronger AI model to train a weaker one, targeting Claude’s most valuable capabilities, including its ability to tackle longer and more complex tasks.
Anthropic urged Congress to penalize companies behind such attacks and ramp up measures to prevent US tech from being stolen, citing the threat to the US military and the potential for Chinese companies to harvest and repackage US AI capabilities as their own.
The company also cited other alleged attacks, which it said posed a threat to the US military, and claimed that such attacks turn hundreds of billions of dollars in American investment into a subsidy for geopolitical competitors.
Implications and Reactions
The allegations have significant implications for the US tech industry, with Anthropic and other US developers accusing Chinese competitors of using distillation attacks to train their models to rival American AI technology at a fraction of the cost.
The incident highlights the growing concern over the theft of US AI technology and the need for increased measures to protect American innovation, with Anthropic’s accusations against Alibaba being the latest example of the ongoing rivalry between US and Chinese tech firms.
