U.S. and China Plan to Compare Pre-Release AI Testing, Bessent Says, While Committing to Nothing That Slows the Race
Treasury Secretary Scott Bessent said on November 10 that next week's Trump-Xi meeting will advance a U.S.-China dialogue on 'superintelligence,' in which the two sides expect to exchange best practices for evaluating and testing advanced models before release and have agreed to a direct channel for alerting each other to serious incidents. Reuters and CNBC report that he made no commitment to limit AI development. No thresholds, agencies or procedures for the channel have been published.
The U.S. and China are building a channel to talk about advanced AI risk without agreeing to slow down. Treasury Secretary Scott Bessent's account of the new "superintelligence" dialogue adds one concrete item, pre-release model testing, to an incident hotline agreed in September, and leaves the competition untouched.
What Bessent said#
In prepared remarks to the National Committee on U.S.-China Relations in New York on November 10, Bessent said President Trump's meeting with President Xi Jinping on the sidelines of the November 18–19 APEC summit in Shenzhen will advance a bilateral dialogue on "superintelligence," or SI, the administration's preferred term, Reuters reported.
"We also expect to exchange best practices for evaluating and testing advanced models before their release, and for keeping human judgment firmly in command of this technology," Bessent said, according to Reuters, adding that the two sides had agreed to establish a direct channel for alerting each other to serious incidents. Reuters says Trump and Xi agreed on the alert system at their White House meeting in September. The dialogue is meant to guard against shared risks and "prevent a misunderstanding between our two systems," he said.
Bessent also said the two countries "should harbor no illusions" about their differences. CNBC reported that he "stopped short of making any commitment to limit advancement of the technology."
What it does and does not establish#
Pre-release testing is the new element. It is also the area where the U.S. administration has preferred voluntary arrangements: in October the White House's Super Intelligence Force said AI incident reporting was "not optional" without issuing a rule. Treasury has separately proposed a ban on U.S. investment in "super intelligence" developers in China and other countries of concern, a reminder that the cooperation track runs alongside restrictions.
Limits#
- Neither government has published what counts as a serious incident, which agencies run the channel, or how testing practices would be shared.
- Bessent's remarks describe expectations for next week's meeting; nothing has yet been signed or announced by both sides.
- China's government has not described the dialogue's terms in the reports we read; China's ambassador spoke at the same event only in general terms, per CNBC.