Donald Trump and Xi Jinping agreed that humans should remain involved in military decisions using AI, US Treasury Secretary Scott Bessent says. If carried through, the reported understanding would draw a line against autonomous decisions to launch major military operations, at a moment when both countries are competing for AI leadership.
Watch Desk analysis
What happened
Bessent told CNBC after returning from Trump’s meeting with Xi in Shenzhen that the leaders had agreed humans should always be involved in military decisions using AI. He said no major military operation should be launched without a person involved. Reuters reported his comments in a story carried by The Straits Times.
Bessent described the understanding as one of the trip’s major achievements. The account establishes what Bessent said about the leaders’ agreement; it does not set out a formal text, how the commitment would be applied, or what enforcement might look like.
Why it matters
A commitment to human involvement sounds straightforward until someone has to define “involved”. The practical difference between a person authorising an operation and a person merely monitoring an automated system is not a footnote when military decisions are at stake.
The reported agreement also puts AI governance alongside the two countries’ competition over technology. Bessent argued that competition would not change the stated understanding. Whether the leaders’ words lead to shared rules or operational safeguards is a separate question, and the report supplies no detail on either.
Our read
This is a consequential statement about the limits leaders say they want on military AI, not evidence of a detailed or enforceable framework. The useful next step is to ask what “human involvement” means in practice: who makes the decision, with what time and information, and whether a person can stop an operation rather than simply watch it unfold.
What to watch
- Whether either government publishes a formal account or text of the understanding.
- How the two sides define human involvement in military decisions using AI.
- Whether the stated principle becomes a practical safeguard, and how compliance would be assessed.
Discussion spark: What should count as meaningful human involvement in a military AI decision: final approval, the ability to stop an operation, or something stronger?
Sources and evidence
- Source update (20 November 2026, 16:50 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.