Beijing’s answer to Amodei’s proposal to slow the pace of development came from the Foreign Ministry on Monday, and it was short. “Fear mongering, confrontation, competition will just disrupt the process of global AI governance,” spokesperson Guo Jiakun said, per an English translation published by Reuters. He had been asked about Amodei, Sam Altman and Elon Musk.
The intelligence minister writes about AI
A day earlier, Chen Yixin, China’s Minister of State Security, published an article calling for accelerated construction of a system to prevent and control AI security risks.
AI, he wrote, has become “the main battleground for global technological competition and a new arena for strategic rivalry among major powers.” What’s needed is “healthy and orderly” development. Two voices out of Beijing, then, in two registers: the Foreign Ministry dismisses the warnings, the security apparatus wants to build up.
The state-owned Global Times was blunter. Amodei’s proposals “seek to portray China’s legitimate development in AI as a threat and further fuel confrontation between China and the US in the field.”
What Amodei wrote about this himself
The point is awkward because Amodei raised it in his own essay. Pacing, he argued, is bounded by the lead U.S. companies hold over “authoritarian regimes, chiefly the Chinese Communist Party”: slow down by more than that margin and unpaced CCP-associated projects pull ahead, creating significant national security risk.
On Sunday he called this the toughest dilemma in his proposal, speaking to CBS News’ “Sunday Morning.” The longer-term answer would be working together to put a speed limit on the rate of AI progress. “I think that’s going to be very difficult because the incentives to pull ahead and the military advantage that you get from that are so large. And honestly, I don’t know if it’s possible, but we should try.”
Six to nine months
Helen Toner, executive director of the Georgetown Center for Security and Emerging Technology and a former OpenAI board member, gave the number on CNBC’s “Squawk Box”: the best estimates put China’s very best models something like six to nine months behind the very best U.S. models.
“If that means that we slow down and they just rip past us because they’re pretty close behind – that’s a problem,” Toner said. She attached a caveat, though: it isn’t clear how much of China’s progress is fast following. If the U.S. frontier slows, China may slow somewhat too.
Amodei’s plan rests on a number nobody has
That’s where the proposal is decided. If Chinese labs do their own research, every brake applied in the U.S. costs lead time. If they mostly follow, braking costs almost nothing – the pace drops on both sides.
Toner says openly that she doesn’t know. Amodei says he doesn’t know either. And the proposal they’re both discussing defines the permitted amount of slowing in terms of exactly that gap. This isn’t a footnote, it’s the central quantity in the plan – estimated from benchmarks rather than measured.
Sources: