Rep. Ro Khanna, ranking member of the House Select Committee on China, is pushing for a binding, enforceable, and verifiable treaty between the United States and China to control the development of artificial intelligence, arguing the alternative is a loss of human control over the technology.
Last month Khanna convened the first congressional hearing calling for such a treaty. He also led every Democrat on the Committee in sending the White House a letter demanding global action on AI risks. Within days, President Trump and General Secretary Xi announced they would establish a bilateral dialogue on AI, which Khanna credits to congressional pressure.
Khanna says that dialogue is only a first step. At his hearing, AI experts called for the U.S. and China to form joint technical working groups of top scientists from both countries to develop verification technologies that would prevent either side from cheating on a future treaty.
Five Pillars
Khanna laid out five elements he says a U.S.-China AI treaty must contain:
- A worldwide pause on recursive self-improvement — AI developing AI without human intervention, which Khanna calls the most likely path to losing control of the technology.
- A global pause on superintelligence, defined as AI that exceeds human capability across all domains.
- Kill switches embedded in all advanced models, though Khanna notes scientists disagree on how effective these would be.
- Mandatory air-gapped sandbox testing of every model, American or Chinese, before release.
- A verification and inspection regime — invoking Reagan’s
