House Panels Launch Investigation Into U.S. Companies' Use of Chinese AI Models
Key Takeaways
- ▸House committees investigating Anysphere's Composer 2 (built on Moonshot AI's Kimi) and Airbnb's use of Alibaba's Qwen model over national security concerns
- ▸Lawmakers cite risks including data access vulnerabilities, supply chain concerns, and potential surveillance risks tied to Chinese law requiring companies to assist intelligence efforts
- ▸Anthropic and OpenAI have previously disclosed evidence of Chinese AI companies (DeepSeek, Moonshot AI, MiniMax) attempting to distill their frontier models through fraudulent accounts and obfuscated methods
Summary
The House Homeland Security Committee and the House Select Committee on the Chinese Communist Party have launched a formal investigation into Anysphere and Airbnb over their reliance on artificial intelligence models developed by Chinese companies. The inquiry specifically targets Anysphere's recently released Composer 2 model, which is built on Kimi, developed by Beijing-based Moonshot AI, and Airbnb's use of Qwen, an AI model created by Alibaba. Republican lawmakers cite national security and cybersecurity risks including potential data access vulnerabilities and concerns tied to a 2017 Chinese law requiring domestic companies to assist state intelligence efforts. The investigation reflects broader mounting concerns among U.S. policymakers about Chinese AI companies allegedly conducting large-scale distillation campaigns to extract knowledge from American frontier models like those developed by OpenAI and Anthropic.
- Investigation seeks details on companies' rationale for using Chinese systems and requests employee participation in in-person briefing with lawmakers
