Anthropic Accuses Alibaba of Illegally Extracting Claude AI Capabilities

Anthropic has accused Chinese tech giant Alibaba of illegally extracting capabilities from its Claude AI model. The case has major implications for intellectual property in the AI industry and the growing technological rivalry between the US and China.

Anthropic, the American AI company behind the language model Claude, has accused Chinese tech giant Alibaba of illegally extracting capabilities from its AI model. The accusation, which became public on June 24, 2026 via Reuters and the BBC, marks a new and striking chapter in the growing tension between Western and Chinese AI companies. The case is drawing worldwide attention and raises fundamental questions about intellectual property in the AI era.

What is 'illegal extraction' of AI capabilities?

The term Anthropic uses — "illicitly extracted" — points to a practice known in the AI industry as model distillation. This involves systematically querying a powerful AI model and then using the responses as training data for a new, competing model. The resulting model effectively 'learns' the knowledge and capabilities of the original, without the creator's permission or any form of compensation. Almost all major AI providers, including Anthropic, explicitly prohibit this in their terms of service. When it happens at scale, companies refer to it as intellectual property theft — and that is precisely what Anthropic is now accusing Alibaba of.

Background: the US-China AI rivalry

Alibaba, through its cloud and AI division, is one of the largest players in the Chinese AI sector and develops, among others, the widely used Qwen series of language models, which are deployed internationally on a large scale. The accusation fits within the broader context of the intensifying technological battle between the United States and China. Anthropic, founded in 2021 by former OpenAI employees including Dario Amodei and Daniela Amodei, positions itself as one of the most serious AI safety companies in the world. Claude competes directly with OpenAI's GPT-4o and Google's Gemini, and is regarded as one of the most advanced commercially available language models. This is not the first time concerns have been raised about the use of outputs from Western models in the development of Chinese AI systems. However, a direct, public accusation aimed at a large, publicly listed company like Alibaba is remarkably bold and unprecedented in its openness.

Legal and ethical implications

If the accusations prove correct, the legal consequences could be far-reaching. Copyright protection for AI models and their outputs remains largely uncharted legal territory, but Anthropic's terms of service are contractually binding for anyone who gains access to the Claude API. A potential lawsuit could set a precedent for the entire AI industry, with the outcome partly determining how companies can protect their models going forward. Experts expect this to also drive stricter technical security measures, such as detecting systematic extraction attempts through anomalous usage patterns in API traffic. Beyond the legal dimension, the case also raises ethical questions. Who owns the knowledge an AI model has acquired? And to what extent should companies be allowed to benefit from the billions of dollars others have invested in building their models?

Consequences for the AI market

The accusation could have significant consequences for how AI companies approach API access and user monitoring. More and more companies are investing in detection mechanisms capable of flagging unusual querying patterns — a telltale sign that someone may be attempting to systematically copy a model. For Alibaba, the reputational stakes are also high. The company operates in an international market where trust and fair use of technology carry increasing weight. A public accusation from a reputable AI safety company could seriously damage Alibaba's standing in Western markets. For companies looking to implement AI applications, this case underscores the importance of carefully selecting trustworthy AI partners and strictly adhering to terms of service.

Conclusion: a pivotal moment for AI intellectual property

Anthropic's accusation against Alibaba marks a significant milestone in the history of artificial intelligence. As AI models become increasingly valuable, so too does the importance of protecting the underlying technology. This case could well become a turning point in how the industry handles intellectual property in the AI era. It is expected that more AI companies will take similar steps in the future to shield their models from unauthorized use. Follow more AI news on stersoftware.com or visit our knowledge base for more background on the latest AI developments worldwide.

BBC News

---

Source: BBC News