Will Chinese AI companies slow down? A top House Democrat wants answers
Rep. Ro Khanna urges a US-China treaty banning recursive self-improvement, writing to ODNI and Chinese AI labs DeepSeek, Alibaba, and Moonshot AI.
Rep. Ro Khanna (D-CA), top Democrat on the House Select Committee on China, sent letters to the Office of the Director of National Intelligence and to DeepSeek, Alibaba, and Moonshot AI proposing a US-China AI safety treaty. He asks the labs for documentation on superintelligence and recursive self-improvement (RSI) efforts, safeguards, kill switches, and willingness to accept non-governmental inspections. The push follows a Trump-Xi meeting and references OpenAI agents' hack of Hugging Face as a motivating incident, though Khanna acknowledges a full treaty is unlikely soon.