Schmidt proposes no-surprise rule for U.S.-China AI cooperation

Eric Schmidt has proposed a no-surprise rule for U.S.-China AI relations, modeled on missile launch notification protocols used by the Pentagon. He argued that rational conversation between U.S. and Chinese elites is necessary to manage the risks of superhuman AI, suggesting that both nations should agree on safety accommodations to prevent catastrophic outcomes. Schmidt noted that while the U.S. has better safety controls and voluntary lab agreements, China’s one-party state structure allows for centralized strategic alignment. He emphasized that AI is not a zero-sum game and that the best outcome involves collaboration to ensure human flourishing, despite current geopolitical tensions.

Schmidt also criticized U.S. immigration policies for expelling talented researchers, citing the case of a Chinese quantum program leader who was kicked out of the U.S. and now leads China's quantum efforts. He contended that unilateral competition is dangerous and that shared rules are essential for managing superintelligent AI risks. Schmidt's proposal aims to establish diplomatic cooperation and safety protocols to mitigate the potential for catastrophic outcomes in the development of advanced artificial intelligence.