Why a Global AI Safety Deal May Be Impossible Without U.S.–China Trust

As fears grow about the misuse of artificial intelligence in warfare, surveillance and critical infrastructure, many policymakers say the world needs something like a global safety compact. Yet a recent report from Hong Kong highlights a hard truth: without a basic level of trust between the United States and China, any serious deal will be difficult to achieve.

U.S. officials increasingly view China’s AI push as a security threat, pointing to its integration into military systems and domestic surveillance. Washington has tightened export controls on advanced chips and restricted investment in Chinese AI firms. Beijing, for its part, accuses the U.S. of trying to contain its technological rise by cutting off access to hardware and cloud services.

Diplomats at the UN and other forums continue to debate AI safety and arms‑control style measures, but observers quoted in the article warn that meaningful, verifiable rules will require Washington and Beijing to agree on shared risks and transparent monitoring – something that looks distant today. The piece compares the situation to climate talks, arguing that the two powers must find a way to “compete and cooperate at the same time.” It also notes that actors like the EU, Japan and middle‑power democracies could play a convening role by pushing for common standards on transparency, incident reporting and testing of high‑risk systems, even if the superpowers remain wary of each other.