中美人工智能安全合作路径
一种更为务实的做法是认识到，人工智能安全合作契合中美两国的共同利益。双方能够拓展关于前沿AI风险的专业对话，强化针对重大AI事件的沟通渠道，促进人工智能评估及安全测试的联合研讨，并寻求在军事AI及网络稳定性方面建立信任的措施。
A more pragmatic approach is to acknowledge that AI safety cooperation aligns with the interests of both nations. China and the U.S. could expand expert discussions on frontier AI risks, improve communication regarding major AI incidents, promote joint discussions on AI assessment and safety testing, and look into confidence-building measures in military AI and cyber stability.
实际上，中美之间已具备一定的合作基础。过去数年，两国专家已通过多项关于人工智能与国际安全的二轨对话展开协作。科学家与研究人员持续就风险治理和新兴技术威胁交换见解。此类交流虽未必成为新闻头条，但其价值在于在技术日益不确定的时代帮助降低误判与误解。鉴于当前许多传统双边互信渠道已明显减弱，人工智能安全仍是少数仍存建设性接触空间的领域之一。
Some foundations already exist. In recent years, Chinese and American experts have collaborated through several Track II dialogues on AI and international security. Scientists and researchers from both sides keep exchanging views on risk governance and emerging technological threats. These exchanges may not make headlines, but they are valuable as they help reduce misunderstanding in a tech landscape that is becoming increasingly uncertain. At a time when traditional bilateral trust channels have weakened, AI safety is still one of the few areas where constructive engagement is possible.
然而，这对华盛顿而言也意味着思维方式的转变。将中国视为人工智能领域的唯一战略对手，虽然能形成国内政治共识，但也可能压缩双方在共同风险问题上的务实合作空间。并非所有人工智能治理议题，都应置于技术遏制或意识形态竞争的框架内去理解。
For Washington, this requires a mindset shift as well. Viewing China solely as a strategic rival in AI may create domestic political consensus, yet it risks limiting the space for practical cooperation on shared risks. Not every AI governance issue should be interpreted through the lens of technological containment or ideological competition.
中美两国真正面临的问题，不仅仅是人工智能能力竞争中的领先与否。更关键的问题是：世界最大的两个大国能否以足够负责任的态度行事，防止人工智能相关风险演变为波及所有人的全球性危机。正因如此，人工智能安全并非中美关系中的边缘议题。在这个充满不确定性的时代，它或许将成为检验两国间是否仍存合作可能性的最重要议题之一。
The core question facing both countries is not just who leads in AI capability. It is whether the world's two largest powers can act responsibly enough to stop AI risks from escalating into a global crisis affecting everyone. That is why AI safety is not a peripheral issue in U.S.-China relations. It may become the most vital test of whether cooperation remains possible in this age of uncertainty.
*peripheral (adj.)次要的、边缘的