Rep. Ro Khanna Urges US-China AI Treaty, Presses Chinese AI Firms on Safety

In letters to Director of National Intelligence Jay Clayton, Khanna asked ODNI to assess how effectively the United States might respond if an AI lab lost control of its bots, and to evaluate the Chinese government’s approach to catastrophic AI risks. He wrote that American and Chinese interests “could not be more different on many issues,” but that “on this discrete concern, it benefits neither country for humanity to lose control of our destiny.” The information, he said, would help Congress and eventually the White House craft an agreement with Chinese counterparts.

The letters to DeepSeek, Alibaba and Moonshot AI ask for documentation about their efforts to achieve “superintelligence” and recursive self-improvement, or RSI, a milestone where AI systems can improve themselves. Khanna also asked about safeguards and “kill switches,” and whether the companies would commit to inspections by a non-governmental agency if a treaty were struck. He asked whether they would be prepared for an incident like OpenAI agents’ hack of Hugging Face this summer. The companies and ODNI did not immediately respond to requests for comment from The Verge.

The move follows a White House meeting last week between Chinese President Xi Jinping and Trump. Khanna said both leaders fail to recognize the risk and urgency around AI. “I think it was constructive that Trump and Xi Jinping at least agreed to talk about AI and have ongoing dialogue about it, but it is not nearly adequate,” Khanna told The Verge in an interview Monday.

Khanna also criticized a planned Tuesday meeting between the White House, House Speaker Mike Johnson (R-La.) and several American tech executives. “I think it’s an effort to reassure the American people that these tech companies are going to be responsible, and it’s a branding exercise,” Khanna said. “But why should we trust companies that have a motive to maximize profits to be making the rules for what is safe? It’s not that they’re bad people, but would you trust the pharmaceutical executives to determine the rules for what drugs are safe?”

Khanna ultimately seeks a U.S.-China treaty banning RSI and monitoring frontier AI labs, but he recognizes that is unlikely to happen soon. Sen. Bernie Sanders (I-Vt.) has similarly called for an AI treaty between the countries. Republicans such as House China Committee Chair John Moolenaar (R-Mich.) have been skeptical the Chinese government would keep its end of the bargain. In the meantime, Khanna said he would like a working group with experts from both countries to suggest safety recommendations and methods for ensuring compliance.

Trump and his allies have been resistant to most regulations or agreements. On the morning of his meeting with Xi last week, Trump posted on Truth Social, “Super Intelligence (SI) will be a big topic of discussion, but I want to leave it exactly where it is. That is China’s position also.” Johnson said on Fox Business that he expected the conversation to revolve around striking “the right balance” of oversight without stifling innovation, warning against losing the AI race against China. Leading U.S. AI companies, however, have called for the United States to coordinate with China to slow down the pace of development.

In the letter to the Chinese AI companies, Khanna acknowledged they are unlikely to respond in full. He said they did not respond to requests in an earlier letter asking for a commitment to slow AI development alongside OpenAI and Anthropic. Khanna told The Verge his goal is to “shame them” and “spark a debate” among Chinese AI companies about safety. While he is critical of U.S. AI firms too, he said, “at least they’re engaging in safety conversations.” Khanna remains confident the United States will drive AI safety standards.

Editor's Summary Rep. Ro Khanna is pressing U.S. intelligence and Chinese AI companies on catastrophic AI risks as he seeks a US-China treaty banning recursive self-improvement and monitoring frontier AI labs. The letters follow a Trump-Xi meeting and precede a White House meeting with tech executives, but Trump and Republican leaders remain resistant to new regulations or agreements.