China and US launch AI talks

China’s Foreign Ministry said on Tuesday that Beijing and Washington had agreed to launch an intergovernmental dialogue on artificial intelligence, adding a new channel of contact to a relationship strained by tariffs, export controls and competition over advanced chips. The announcement came five days after U.S. (news.cgtn.com) President Donald Trump and Chinese President Xi Jinping met in Beijing, where artificial intelligence was discussed alongside broader economic and security issues. Chinese foreign ministry spokesperson Guo Jiakun said the two countries should work together on AI development and governance. (english.aawsat.com) U.S. Treasury Secretary Scott Bessent said on May 14 that the two sides would establish a protocol for AI safety. (cnbc.com) ### Where did this come from? May 19 was the first public confirmation from Beijing that the two governments had agreed to start a formal AI dialogue. Guo made the announcement at a regular foreign ministry briefing, according to Chinese state media and other reports citing the ministry. May 14 was the first public signal from Washington that such talks were being set up. Bessent told CNBC from Beijing that the United States and China would establish a protocol for artificial intelligence safety and said the conversation was possible because, in his words, the United States was “in the lead.” (news.cgtn.com) ### What are the two sides saying the talks will cover? Guo said China and the United States should work together to promote both the development and governance of AI. The phrasing from Beijing puts development and governance side by side, rather than limiting the channel to security risks alone. (news.cgtn.com) China’s published AI governance documents show the line Beijing has been advancing in multilateral settings. A foreign ministry initiative first released in 2023 calls for countries to enhance information exchange and technological cooperation, prevent risks, and develop AI governance frameworks, norms and standards based on broad consensus. (cnbc.com) Bessent described the U.S. side more narrowly as an effort to establish safety best practices. Reuters reported on May 14 that U.S. and Chinese delegations were discussing guardrails for the most powerful AI models and a protocol to keep non-state actors from exploiting them. (english.aawsat.com) ### Is this a brand-new channel or a revival? The United States and China had already opened formal AI talks under former President Joe Biden and Xi at their 2023 California summit, according to reporting that cited people familiar with the earlier process. Those talks also produced a 2024 understanding that humans, not AI, would retain authority over nuclear weapons use. (mfa.gov.cn) The current effort appears to revive that contact in a different political setting. Reuters reported on May 6 that Washington and Beijing were weighing official discussions on AI ahead of the Trump-Xi summit in Beijing. (msn.com) ### Why is AI even on the agenda when the tech fight is so broad? May 2026 has brought no easing in the wider technology rivalry. Reuters reported last week that distrust and competition over chips and self-sufficiency were clouding hopes for a broader AI push at the Trump-Xi summit. State Department materials show Washington is also building its own AI diplomacy track with allies and partners. (theoutpost.ai) The department said on May 12 that a U.S. delegation was attending APEC meetings in Shanghai and Suzhou, and its policy pages describe separate work on AI research, development and governance. (usnews.com) China, for its part, has continued to promote international AI governance frameworks through its foreign ministry. Its published initiative calls for open cooperation while also stressing national sovereignty and opposition to using AI to interfere in other countries’ internal affairs. (usnews.com) ### What happens next? May 19’s announcement did not include a date, venue or list of officials for the first round of talks. Beijing said only that an intergovernmental dialogue would be launched. The next concrete marker is likely to be a named protocol, meeting date or lead negotiator from either Treasury, the State Department or China’s foreign ministry. (state.gov) Bessent said on May 14 that the two countries would set up an AI safety protocol, but neither government had published full terms by Tuesday. (cnbc.com) (news.cgtn.com) (mfa.gov.cn)