# Khanna Seeks Binding U.S.-China Agreement on Advanced AI Safety By Riza Dagoc Canonical URL: https://www.tokenpost.com/news/regulation/25321 Published: 2026-09-29T12:31:25.000Z Updated: 2026-09-29T12:31:25.000Z Section: Regulation > The proposal would restrict recursive self-improvement and create international monitoring as Washington and Beijing prepare their first Super Intelligence Dialogue exchange by November 2026. Rep. Ro Khanna, the top Democrat on the House Select Committee on China, is seeking a binding U.S.-China agreement to restrict recursive self-improvement in advanced artificial intelligence systems as Washington and Beijing establish a new channel for discussing superintelligence risks. The proposed agreement would include international monitoring, inspections and verification measures, Khanna said. He also called for safeguards against using AI in biological and nuclear weapons. The United States and China agreed to establish a Super Intelligence Dialogue and a bilateral communication channel for incidents involving superintelligence. The first exchange is due by November 2026. President Donald Trump and Chinese President Xi Jinping also agreed to use the term “super intelligence” instead of “artificial intelligence” for applicable emerging technologies. Khanna compared his proposal with arms-control frameworks that rely on inspection and verification. “Winning an AI war is like winning a nuclear war. It is catastrophic for civilization,” Khanna said. He said Chinese AI companies DeepSeek, Alibaba and other firms had not responded to his request that they slow development until safety guardrails are established, following similar steps by U.S. companies. The proposal extends the policy debate beyond voluntary corporate restraint. Khanna said international monitoring should cover the infrastructure used to develop advanced AI systems and help determine whether companies can maintain control over increasingly capable models. The announced U.S.-China dialogue is narrower than Khanna’s proposal. It establishes communication on superintelligence and related incidents but does not create a binding treaty, mandate a development pause or require inspections. The initiative follows [earlier U.S.-China discussions over AI cooperation and technology restrictions](), as both countries continue to debate how advanced AI should be developed and governed. The first Super Intelligence Dialogue exchange is due by November 2026. ## Links in this article - [earlier U.S.-China discussions over AI cooperation and technology restrictions](https://www.tokenpost.com/news/regulation/23606)