Alibaba Blocks Employee Access to Anthropic AI Amid Corporate Espionage Allegations

Alibaba Group will prohibit its employees from using Anthropic's artificial intelligence tools for work starting July 10, citing potential security risks from the U.S. company's software. The Chinese e-commerce giant has classified Anthropic's coding assistant, Claude Code, as a high-risk application, according to internal sources who spoke on condition of anonymity.

The restriction follows a dramatic escalation between the two companies. In June, Anthropic sent a letter to a U.S. Senate committee accusing Alibaba of orchestrating "the largest known distillation attack" to steal its AI capabilities. Anthropic's terms of service already bar Chinese companies and other "adversarial nations" from using its models. Alibaba employees must now uninstall all Anthropic tools and switch to the company's own AI assistant, Qoder. Both firms declined to comment.

The ban comes as Chinese social media erupts with complaints about Anthropic embedding hidden code to detect users from China. Meanwhile, reports indicate other Chinese tech giants have found workarounds. Ant Group, Alibaba's fintech affiliate, reportedly provided employees with Claude accounts accessed through Singapore-based servers. ByteDance, TikTok's parent, reimburses engineers for personal AI subscriptions they use via VPNs - a policy meant to encourage skill development, according to an insider. Ant and ByteDance declined to comment on the Financial Times report.

Source: CNBC