Alibaba Blacklists Anthropic's Claude Code Over Distillation Attack Fears
Alibaba has added Anthropic's Claude Code to its high-risk software list, restricting employee access amid accusations of a distillation attack.
Chinese e-commerce giant Alibaba has placed Anthropic's Claude Code on an internal high-risk software list, effectively banning employees from using the AI coding tool, according to a report from US Top News and Analysis. The move signals a dramatic escalation in tensions between Chinese tech firms and American AI developers over the security and integrity of proprietary model training data.
The ban centers on allegations of a so-called distillation attack — a technique in which outputs from one AI model are systematically harvested and used to train or fine-tune a competing model, potentially allowing a rival to replicate capabilities without independently developing them. While the source does not specify the precise technical details of the alleged attack, the accusation itself reflects growing anxiety across the AI industry about intellectual property protection and model security.
Read more Vertex Pharmaceuticals to Buy Crinetics in $10B Deal →
Alibaba's decision to formally categorize Claude Code as high-risk software suggests the company views the tool as a potential threat to its own AI development operations, including its Qwen family of large language models. The move also underscores the deepening competitive and geopolitical fault lines between US and Chinese AI ecosystems, where trust between companies on opposite sides of the Pacific is increasingly fragile.
For Anthropic, the blacklisting represents a significant reputational and market-access setback in China, one of the world's largest technology markets. The episode also raises broader questions about how AI companies can safeguard their models against extraction-based attacks as coding assistants become ubiquitous enterprise tools. Industry analysts are likely to watch whether other Chinese firms follow Alibaba's lead in restricting Western AI software on similar grounds.
Continue reading at US Top News and Analysis.
