US-China AI Treaty Proposal
Rep. Ro Khanna (D-CA) is calling for a formal treaty between the US and China to keep advanced artificial intelligence from causing severe global risks. Wrangling leaders in both nations to take concrete action remains a difficult long-term challenge.
Will Chinese AI companies slow down? A top House Democrat wants answers
He states that both American and Chinese leaders lack the necessary urgency to rein in the risks associated with rapid technological growth.
Inquiries Sent to Chinese Labs
In recent letters, Khanna — the top Democrat on the House Select Committee on China — asked a US intelligence agency and multiple Chinese AI companies whether they are prepared for major autonomous security incidents. The contacted organizations include DeepSeek, Alibaba, and Moonshot AI, all of which develop high-level artificial intelligence models.
American and Chinese interests differ fundamentally on numerous issues, Khanna noted in correspondence with Director of National Intelligence Jay Clayton, but on this specific concern, it benefits neither nation for humanity to lose control of its destiny.
“[I]t benefits neither country for humanity to lose control of our destiny”
Khanna requested that the Office of the Director of National Intelligence assess how effectively the US might respond if an AI lab loses control of its autonomous systems. He also asked the agency to evaluate the Chinese government's approach to assessing catastrophic AI risks to help lawmakers craft an eventual international framework.
Letters sent to DeepSeek, Alibaba, and Moonshot AI requested documentation regarding their efforts to achieve superintelligence and recursive self-improvement, a milestone where systems can independently upgrade themselves. The inquiries also address potential safety safeguards, emergency kill switches, and willingness to submit to non-governmental inspections under a potential treaty.
Diplomatic and Industry Reactions
The legislative move follows a high-level meeting between Chinese President Xi Jinping and US President Donald Trump in Washington. Khanna argues that both leaders fail to recognize the immense urgency surrounding frontier artificial intelligence risks.
Khanna expressed skepticism regarding upcoming discussions between the White House, congressional leaders, and major tech executives, characterizing them primarily as corporate branding exercises rather than rigorous safety rulemaking.
“Why should we trust companies that have a motive to maximize profits to be making the rules for what is safe?”
While Khanna's ultimate goal is a bilateral treaty banning recursive self-improvement and monitoring frontier labs, he acknowledges such an agreement is unlikely in the immediate future. Other lawmakers, including Sen. Bernie Sanders, have voiced support for similar diplomatic measures, while Republican critics remain skeptical of compliance from Beijing.
The Trump administration has generally resisted heavy regulatory agreements, with officials emphasizing the need to maintain a competitive edge against foreign rivals without stifling domestic innovation.
Accountability and Technical Expertise
Khanna acknowledged that Chinese AI firms are unlikely to respond comprehensively, noting that previous inquiries yielded no commitments to slow down development. However, he stated that the initiative aims to spark critical safety debates within the industry.
Khanna remains confident that technical experts and engineers, rather than corporate executives or politicians lacking specialized knowledge, must ultimately drive safety standards and policy debates.
