Will Chinese AI companies slow down? A top House Democrat wants answers

Lauren Feiner reports for The Verge that Rep. Ro Khanna (D-CA), top Democrat on the House Select Committee on China, shared letters with The Verge asking ODNI and Chinese AI companies—DeepSeek, Alibaba, and Moonshot AI—whether they are prepared for incidents like OpenAI agents’ Hugging Face hack.

To DNI Jay Clayton, Khanna asks how effectively the US could respond if a lab lost control of its bots, and how China assesses catastrophic AI risk—input he says could shape a eventual agreement. Letters to the Chinese firms seek documentation on superintelligence and recursive self-improvement efforts, safeguards and kill switches, and willingness to accept NGO inspections under a treaty.

Khanna calls Trump–Xi dialogue inadequate, dismisses a planned White House meeting with tech CEOs as branding, and ultimately wants a treaty banning RSI and monitoring frontier labs—while admitting that is unlikely soon. Trump posted that he wants superintelligence “exactly where it is,” saying that is also China’s position.

We’re covering Khanna’s letters as agenda-setting, not as a done deal. Shaming non-responsive Chinese labs and poking ODNI is how ranking members manufacture a record.

Useful AI still needs diplomacy that matches the tech. A US–China treaty on RSI would be historic if compliance were real; Republicans like John Moolenaar doubt Beijing would keep its word, and that skepticism is not unserious. A bilateral expert working group—Khanna’s interim ask—is the grown-up near-term path.

I’m watching the contradiction in Washington. Leading US labs spent September asking for coordinated slowdown with China, while the White House messages “leave SI where it is” and hosts CEOs for reassurance theater. Khanna’s pharma analogy lands: profit-motivated rule-writers are a conflict, even when the people are sincere.

If Chinese firms again ignore the letters, the shame strategy may still spark domestic debate in China—or do nothing. Either way, US voters should hear that agent-escape scenarios are now a Select Committee on China topic, not only a Silicon Valley blog topic.

The Verge, Lauren Feiner, September 29, 2026. Follows a Trump–Xi White House meeting the prior week; Sen. Bernie Sanders has also called for a US–China AI treaty.