Anthropic Accuses Alibaba of Illicit AI Model Extraction
US artificial intelligence (AI) leader Anthropic has accused the Chinese e-commerce and technology giant Alibaba of "brazenly" and "illicitly" extracting capabilities from its Claude AI model.
In a letter reviewed by the BBC, the San Francisco-based company stated that operators linked to Alibaba conducted nearly 29 million interactions with Claude using thousands of fraudulent accounts. Anthropic described this as the largest extraction campaign of its kind.
The company has called on the US Congress to impose penalties on firms responsible for such attacks and to strengthen protections against the theft of US technology.
The BBC has reached out to Alibaba for comment and requested further details from Anthropic.
Details of the Alleged Extraction Campaign
Anthropic's letter, dated 10 June and addressed to US Senators Tim Scott and Elizabeth Warren, accused Alibaba, listed on the New York Stock Exchange, of orchestrating "the largest campaign to illicitly extract Claude's capabilities."
The campaign reportedly employed "distillation attacks," a technique that extracts responses from a more advanced AI model to train a less capable one.
According to Anthropic, Alibaba-linked operators targeted Claude's most valuable features, including its capacity to handle longer and more complex tasks and its decision-making processes.
The company stated these attacks are conducted on an "industrial scale" to allow Chinese firms to appropriate and repackage US AI capabilities as their own.
Broader Security Concerns and Geopolitical Implications
The letter also referenced other alleged attacks that Anthropic claims pose a threat to US military interests.
"Distillation attacks turn hundreds of billions of dollars in American investment and [research and development] into a massive subsidy for our geopolitical competitors,"
Anthropic cited claims from the US Department of Defense linking Alibaba and other major companies such as car manufacturer BYD and tech firm Baidu to the Chinese military.
These companies have denied such allegations. Meanwhile, Alibaba has recently filed a lawsuit against the US government seeking removal from the Pentagon's blacklist.
US developers have previously accused Chinese competitors of using distillation attacks to train their AI models to compete with American technology at significantly lower costs. OpenAI has also made similar accusations against Chinese groups.
Anthropic's Position in the AI Industry
Anthropic is a prominent AI developer and, alongside ChatGPT creator OpenAI, is preparing for a major stock market debut that could position it among the world's most valuable companies.
However, some of Anthropic's advanced models, such as Mythos, have raised cybersecurity concerns due to their potential to identify vulnerabilities in computer systems.
- Is China quietly winning the AI race?
- The Chinese AI app sending Hollywood into a panic
