The United States and China are set to meet in Washington next week, a timing that coincides with what many observers describe as a pivotal moment for artificial intelligence. In the two weeks preceding the summit, a series of high-profile alerts highlighted the technology’s potential to outpace human oversight, prompting calls for coordinated policy action.
One of the most dramatic signals came when Jacob Coxon, a researcher at Anthropic, quit his position and posted that the engineers building AI “earnestly believe that it could kill us all by the end of the decade.” Shortly thereafter, Anthropic’s chief executive Dario Amodei released an essay urging a slowdown of AI progress, warning that unchecked development could exceed our capacity to understand or control it. The proposal received public backing from OpenAI’s Sam Altman and Grok founder Elon Musk.
Chinese experts appear less unified on the existential dimension of the risk. According to the Sydney Morning Herald’s China correspondent Lisa Visentin, Beijing does not frame AI as a threat to humanity in the same language used in Silicon Valley, though some senior officials have expressed unease about possible dangers. She notes that the dominance of the United States and China in the field leaves most other nations watching and hoping that the two powers can reach a workable solution.
Previous diplomatic attempts have yielded limited progress. In May, President Trump and President Xi discussed launching an intergovernmental AI dialogue, but the initiative stalled. Trump later claimed that a “SICK conspiracy” targets AI and data centres and suggested that only China benefits from any regulatory push, while Xi may view U.S. calls to “pace the frontier” as an effort to contain Chinese advancement.
The Washington gathering therefore represents a narrow window for translating alarm into policy. If the leaders can move beyond suspicion and craft a bilateral framework, it could establish the first concrete guardrails for a technology many now deem the most consequential of our era. Failure to act, however, would leave the global community dependent on the goodwill of the two AI superpowers.