President Trump returned from his China trip with a critical agreement on military artificial intelligence that mandates human oversight over systems capable of triggering war or peace.
Treasury Secretary Scott Bessent confirmed that during discussions with Chinese President Xi Jinping, the two leaders established joint guardrails requiring humans to remain involved in all major military decisions made by superintelligent systems.
The principle states that while advanced artificial intelligence is increasingly deployed for intelligence, logistics, cyber defense, and strategic planning, humans must always be the final decision-makers for military actions. Superintelligence systems are explicitly prohibited from independently launching or directing major military operations.
Bessent emphasized that this agreement represents an immediate policy framework rather than a formal treaty, designed to prevent autonomous military choices by artificial intelligence during crises without diminishing the ongoing technological competition between the United States and China.
The White House highlighted that both nations continue their race for advanced technologies while maintaining human control in high-stakes scenarios. Specific implementation details remain to be finalized, including definitions of “major military actions” and verification mechanisms for ensuring human oversight.
This development follows President Trump’s meetings with Chinese President Xi Jinping and Vladimir Putin during his Asia trip.