Scott Bessent says Trump and Xi agreed that military AI decisions should involve humans, and that the spying allegations involving Alice Huang didn't come up
First reported by Reuters ·
The U.S. and China have agreed that human oversight is required for AI in military decisions.
Treasury Secretary Scott Bessent stated that former Nvidia CEO Jensen Huang was not discussed during President Trump's meeting with Chinese President Xi Jinping. Huang resigned amid allegations of extramarital affairs and of funneling restricted chips to China. Bessent emphasized the U.S. need to maintain its lead in AI innovation over China, viewing domestic AI development as a shield against foreign intelligence threats. He also revealed that Trump and Xi agreed on the necessity of human oversight for any military decisions made by AI, prohibiting autonomous AI actions in major military engagements. Bessent reiterated that advancing U.S. AI companies at the frontier is the primary defense against AI-enabled cyberattacks.
The consensus between the U.S. and China on human oversight for military AI suggests a shared concern over escalating AI capabilities and potential unintended consequences in warfare. This alignment could subtly influence the geopolitical landscape of AI development, potentially leading to de-escalation in certain military AI applications or, conversely, a race to develop AI within these agreed-upon human-in-the-loop frameworks. It signals a pragmatic approach from both superpowers, acknowledging the inherent risks of fully autonomous military AI.
Furthermore, Bessent's emphasis on U.S. AI companies maintaining a "substantial lead" and having "access to the best chips" highlights the ongoing strategic importance of semiconductor supply chains in the AI race. This underscores the competitive dynamic, where technological advancement is seen as critical national security. The administration's view of AI companies as a "shield" against foreign threats indicates a strategy of fostering domestic innovation as a primary defense mechanism against state-sponsored cyber operations and espionage.
AI-written summary. May contain errors.