Anthropic Accuses Alibaba of ‘Distillation Attack’ on Claude AI Models Anthropic’s clash with Alibaba over alleged AI model theft has escalated into a test case for how far companies – and governments – will go to protect cutting‑edge artificial intelligence.

Early concerns and context

Anthropic, maker of the Claude chatbot, has been warning for months that rival firms could “extract” advanced model capabilities without paying for the costly research and training required. Alibaba, meanwhile, has been building out its own large language models under the Alibaba Cloud banner, including its Qwen series, as it seeks to compete in the global AI race.

The alleged “distillation attack”

In a June 10 letter to US Senators Tim Scott and Elizabeth Warren, Anthropic’s head of policy Sarah Heck accused Alibaba‑affiliated operators of carrying out “the largest known distillation attack” on its systems. Between April 22 and June 5, those operators allegedly conducted “28.8 million exchanges with Claude through almost 25,000 fraudulent accounts,” Heck wrote, with the goal to “illicitly extract Claude’s capabilities” to train Alibaba’s own models.

Heck described these distillation attacks as being conducted “illicitly, systematically, and at industrial scale to harvest US AI capabilities across frontier labs and repackage them as their own without incurring the training and R&D costs required.” Anthropic has framed this not only as a commercial threat but as a national‑security issue, warning that such tactics could help Chinese models rapidly approach the capabilities of Claude’s most advanced variants.

Regulatory and geopolitical fallout

Anthropic’s letter landed just weeks after the US government imposed export controls on its latest Fable 5 model, barring foreign individuals from accessing it on national‑security grounds. Around the same time, Alibaba was added to a Pentagon blacklist linking certain companies to the Chinese military – a designation the firm is now challenging in US court.

Anthropic is urging lawmakers to respond with new rules targeting distillation attacks, including tighter limits on China’s access to advanced US computing infrastructure and penalties for entities found to be engaging in such practices. The dispute underscores a broader tension: whether open access to powerful AI can coexist with intensifying geopolitical competition over who controls the technology.

Continue reading https://foxvector.com/stories/019efe86-f805-1085-73b1-14e4298d2193