US China plan September AI safety talks

The United States and China are planning formal bilateral AI talks for September, marking the first such dialogue under President Donald Trump, with Treasury Secretary Scott Bessent leading the American delegation to address risks posed by frontier AI models and the potential spread of advanced tech to non-state actors. The talks stem from the May Trump–Xi summit and are expected to occur before Xi Jinping’s visit to the United States on September 24, though exact dates and venues remain undecided as preparations proceed. While the agenda is still being finalized, officials say the discussions will cover AI safety and governance, export controls, and how to manage military applications and cyber threats. The effort signals a push to formalize communication amid intensifying U.S.-China tech rivalry and to shape norms around powerful AI, with Washington holding a perceived lead in the talks. Chinese and U.S. policymakers are also weighing regulatory responses to open-source models and broader restrictions as part of this strategic dialogue.