Anthropic Accuses Alibaba of Illicitly Extracting AI Capabilities

Jun 25, 2026

Anthropic has accused Alibaba of launching a massive distillation attack, using 25,000 fraudulent accounts to conduct 28.8 million exchanges with its Claude model to extract advanced capabilities. The incident highlights escalating US-China AI tensions, as the Trump administration restricts Anthropic's powerful Mythos model over national security concerns, while Chinese tech leaders urge domestic development of rival cybersecurity AI to counter American dominance.

Alibaba distillation attack on Claude

▪Operators affiliated with Alibaba and its Qwen AI lab conducted approximately 28.8 million exchanges with Claude using about 25,000 fraudulent accounts between April 22 and June 5, 2026.
▪Alibaba used obfuscation techniques and proxy networks to evade detection during the distillation campaign.
▪The distillation campaign targeted Claude's advanced capabilities, including agentic reasoning, software engineering, and long-horizon tasks.
▪Anthropic sent a letter on June 10, 2026, to U.S. Senators Tim Scott and Elizabeth Warren accusing Alibaba of conducting a massive distillation attack on its Claude AI model.

US-China AI competition

▪A flood of cheap, powerful Chinese AI models, such as z.AI's GLM-5.2 released on June 16, 2026, are gaining U.S. customers and challenging American AI leadership.
▪The White House Office of Science and Technology Policy issued a memorandum in April 2026 pledging to assist AI companies in detecting and coordinating against mass distillation.
▪Anthropic stated that distillation attacks turn billions of dollars in American research and development into subsidies for geopolitical competitors like China.

Mythos cybersecurity capabilities race

▪Anthropic's Mythos model is capable of detecting software vulnerabilities and outperforming humans on cybersecurity tasks, which Chinese tech figures have likened to a strategic cyber weapon.
▪Zhou Hongyi, founder of 360 Security Technology, warned that China must develop its own Mythos-like model to avoid a significant disadvantage in cyber warfare.

Trump administration enforcement response

▪The U.S. Commerce Department held off on placing Chinese AI startup DeepSeek on a trade blacklist despite it being deemed a national security risk.
▪The Trump administration issued an export control directive on June 12, 2026, ordering Anthropic to suspend global access to its Fable 5 and Mythos 5 models due to national security concerns.
▪The U.S. Department of Defense added Alibaba to its list of Chinese military companies in June 2026, a designation Alibaba challenged in a lawsuit filed on June 23, 2026.

Anthropic legislative recommendations

▪Anthropic recommended that Congress update antitrust laws to allow AI firms to share threat intelligence regarding Chinese distillation tactics.
▪Anthropic urged Congress to implement stricter export controls on chips and penalize Chinese labs by restricting their access to U.S. models, chips, or data centers.

Chinese cybersecurity industry reaction

▪Alibaba's stock price fell by 3% to 4% following the public release of Anthropic's distillation accusations.
▪Zhou Hongyi stated that Chinese firms are currently well short of Mythos-level capabilities and are focusing on AI agent systems combining foundation models with specialist security data.