Did Alibaba Illegally Extract Anthropic’s AI Capabilities?

2026-06-25 · Source: AI Magazine · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cybersecurity & Data Privacy · Depth: Fundamental Awareness, short

Summary

Anthropic has accused Chinese e-commerce and technology giant Alibaba of orchestrating a large-scale "model extraction" campaign to illicitly acquire capabilities from its Claude AI models. The AI startup claims Alibaba used approximately 25,000 fraudulent accounts to generate over 28.8 million exchanges with Claude between April 22 and June 5, 2026, aiming to replicate the performance of Claude Mythos Preview. This alleged intellectual property infringement utilizes knowledge distillation, a legitimate machine learning technique, to train "student models" from proprietary "teacher models." Anthropic, which sent a letter detailing these claims to U.S. Senators Tim Scott and Elizabeth Warren on June 10, 2025, has previously made similar accusations against other Chinese AI firms like DeepSeek, Moonshot AI, and MiniMax, highlighting an intensifying US-China race for AI dominance.

Key takeaway

For policy makers and legal professionals assessing AI intellectual property protection, this incident underscores the urgent need for robust frameworks against model extraction. Your current IP laws may be insufficient to deter sophisticated, large-scale distillation campaigns that replicate proprietary AI capabilities. Consider developing international agreements and domestic regulations specifically addressing AI model theft. Mandate closer collaboration between AI firms and government agencies to counter these evolving threats.

Key insights

Model extraction attacks exploit knowledge distillation to illicitly replicate proprietary AI model capabilities, raising significant IP concerns.

Principles

• Knowledge distillation can be weaponized for AI intellectual property theft.
• Countering sophisticated model extraction requires industry-government coordination.
• The US-China AI race intensifies with accusations of capability siphoning.

Method

Threat actors generate vast input-output pairs from proprietary LLMs using fake accounts, then use these to train "student models" that mimic the "teacher model's" performance.

Topics

• Model Extraction
• Knowledge Distillation
• AI Intellectual Property
• US-China AI Competition
• Anthropic Claude
• Alibaba Qwen

Best for: CTO, VP of Engineering/Data, Director of AI/ML, Legal Professional, Tech Journalist, Policy Maker

Related on AIssential

• Anthropic Paid $1.5B for Pirated Data — Frontier AI labs are reporting coordinated, large-scale distillation attacks by foreign entities to extract model capabilities.
• Anthropic accuses DeepSeek, Moonshot, and MiniMax of "industrial-scale distillation attacks". — AI development faces critical challenges in intellectual property protection, ethical data use, and robust model evaluation amidst rapid technological advancements.
• Breaking 🚨: Anthropic vs China! — Anthropic accuses Chinese LLM firms of illicitly distilling Claude's capabilities, framing it as a geopolitical threat.
• Trump science advisor says Chinese actors are copying American AI at massive scale — China is allegedly using large-scale distillation to copy US AI models, stripping safety features in the process.
• Anthropic says Alibaba must be punished for largest Claude cloning attack — Chinese firms are allegedly conducting large-scale distillation attacks to replicate US frontier AI capabilities, raising national security concerns.
• Anthropic takes U.S. government to court — Anthropic is suing the U.S. government, alleging retaliation for its AI safety advocacy.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Magazine.