Beijing dismissed calls from US tech leaders to slow down AI development as self-serving fearmongering.
It’s not that Chinese officials don’t want to discuss guardrails; they do — but not on terms set by the US.
Artificial intelligence is expected to feature prominently when presidents Donald Trump and Xi Jinping sit down in Washington in less than two weeks.
One reason that China’s pushing back is that it still lags behind the US, and Beijing sees AI as a once-in-a-lifetime chance to reshape the tech balance in its favor.
There’s also huge skepticism as to the underlying intention of American calls to hit the brakes.
Chinese AI companies already face US blocks on the chips needed to train and operate frontier models. State-controlled China Daily accused the US AI executives of wanting to further blunt Chinese competition.
It doesn’t help that Anthropic’s Dario Amodei is out in front.
Beijing holds special antipathy for the American entrepreneur, who has emerged as a prominent advocate of tighter US restrictions on China’s access to advanced AI.
Yet safety concerns are also front of mind in China, even if areas Beijing considers important don’t always fully overlap with American worries.
One prominent fear is that AI could threaten the legitimacy of the Communist Party, as voiced in an article posted Sunday by the country’s top intelligence official.
China’s definition of AI security is broader and continues to expand.
Just yesterday it released an updated safety framework to address emerging risks, including those of “recursive self-improvement” — the same dynamic cited by Amodei arguing for greater caution.
Even as Trump downplays the need for guardrails, the US and China have shared concerns, from AI-enabled cyberattacks by non-state actors to keeping humans in control of nuclear decisions.
Amid all the AI-focused noise, perhaps that’s something worth remembering as the Xi-Trump summit approaches. — Colum Murphy