Artificial intelligence firm Anthropic has accused Chinese technology giant Alibaba of orchestrating a large-scale operation to extract capabilities from its Claude AI models.
Anthropic accused Alibaba of using nearly 25,000 fraudulent accounts, escalating concerns over intellectual property protection in the rapidly evolving AI industry.
The allegations, outlined in a letter sent to US legislators and White House officials, claim that operators linked to Alibaba's AI division, Qwen, conducted approximately 28.8 million interactions with Claude between April and June this year.
Anthropic says the activity was designed to systematically study and replicate the behaviour of its advanced AI systems.
According to the company, the campaign focused on extracting Claude's strengths in software engineering, complex reasoning, long-term planning and advanced problem-solving, capabilities considered highly valuable in the race to develop increasingly powerful AI models.
At the centre of the dispute is a technique known as "AI distillation."
In the AI industry, distillation refers to the process of training a smaller model using outputs generated by a more advanced model. The approach is widely used to improve efficiency and reduce computing costs.
However, Anthropic argues that Alibaba's alleged activities went beyond accepted industry practices.
The company claims the use of thousands of fake accounts allowed operators to repeatedly query Claude at scale, generating massive amounts of data that could potentially be used to train competing AI systems.
Anthropic described the operation as being conducted "illicitly, systematically and at an industrial scale."
The accusations arrive amid growing competition between American and Chinese technology companies seeking dominance in artificial intelligence.
Developing frontier AI models requires billions of dollars in investment, access to advanced computing infrastructure and years of research.
As a result, companies increasingly view their AI systems as strategic assets deserving the same protection as patents, trade secrets and proprietary software.
Industry analysts note that if allegations of large-scale AI extraction become commonplace, developers may be forced to introduce stricter safeguards to prevent competitors from harvesting model outputs.
Anthropic has urged US policymakers to strengthen protections against unauthorised AI distillation and improve mechanisms for detecting coordinated extraction attempts.
Alibaba is yet to respond
At the time of publication, Alibaba had not publicly issued a detailed response to the allegations.
The company has reportedly declined or not immediately responded to requests for comment from several international media outlets.
Importantly, no court or regulatory body has independently verified Anthropic's claims, and the allegations remain unproven.
News of the accusations triggered an immediate reaction in financial markets, with Alibaba shares falling as investors weighed the potential regulatory and reputational implications.
The dispute also highlights the increasing intersection between artificial intelligence, international competition and national security, as governments seek to protect critical technologies from unauthorised access.
The controversy underscores a growing challenge facing the AI industry: determining where legitimate research ends and intellectual property infringement begins.
As AI models become more sophisticated and expensive to develop, companies are expected to invest heavily in protecting their technologies from unauthorised replication.
The outcome of disputes such as the one between Anthropic and Alibaba could shape future regulations governing AI development, model training and digital intellectual property rights.
Comments
Sign in with Google to comment, reply, and like comments.
Continue with Google