ROSS DOUTHAT2026年7月21日Early in the Cold War, even as the United States and the Soviet Union engaged in a relentless nuclear arms race, they treated nuclear technology as a tool of soft power and diplomacy — something to share and export, a seal on alliances and a token of friendship.冷战初期，就在美苏两国进行无休止的核军备竞赛的同时，双方也将核技术视为软实力与外交的工具——一种可以共享和出口的资源，是巩固同盟、表达友谊的信物。Thus Dwight Eisenhower’s America launched the Atoms for Peace program, which helped build civilian nuclear programs around the world, in the same year that it executed the Rosenbergs for giving the Soviets our atomic secrets. The Soviet Union likewise exported forms of nuclear power to Eastern Bloc nations and Mao’s China. In each case, the broad goal was to spread the tech while maintaining control over the weapons, taking very different approaches to the two forms of nuclear strength.因此，正是在艾森豪威尔政府处决向苏联泄露美国核机密的罗森堡夫妇的同一年，美国启动了“原子用于和平”计划，帮助世界各国建立民用核项目；苏联也向东欧的社会主义阵营国家以及毛泽东时代的中国输出了各种形式的核能技术。两国的总体目标都是在保持对核武器控制权的同时传播核技术，只是对于核力量的两种不同形式采取了截然不同的策略。This history suggests a framework for understanding the difference between the American and Chinese approaches to artificial intelligence. The United States is behaving as though frontier A.I. models could be the equivalent of nuclear weapons, offering an immensely destructive form of power that requires a careful and jealous watch. American A.I. tycoons constantly talk as if they were engaged in an Oppenheimer-style race (complete with Oppenheimer-style moral hand-wringing), the American military-industrial complex is intent on keeping China from closing the current A.I. gap, and both industry and government increasingly treat frontier models as grave security threats.这段历史为我们提供了一个理解美中两国人工智能路径之异同的框架。美国的做法仿佛把最前沿的人工智能模型视为核武器——一种具有巨大破坏力、需要小心翼翼、严密守护的力量。美国的人工智能巨头们时刻挂在嘴边的语气仿佛他们正置身于一场奥本海默式的竞赛（连同奥本海默式的道德焦虑），美国军工复合体一心阻止中国缩小当前的人工智能差距，产业界和政府也越来越倾向于将前沿模型视为严峻的安全威胁。Meanwhile, China treats the models that its companies produce as the equivalent of nuclear energy, a lower-risk technology to be shared and commercialized to win friends and influence the world and undermine the advantages of the American A.I. giants. There is some Chinese lip service to risk and safety, but little sense that this is a unique power to be hoarded, a geopolitical trump card or a weapon with existential stakes.与此同时，中国则把本国企业开发的人工智能模型视作核能——一种风险较低的技术，应当通过共享和商业化推广来赢得朋友、扩大国际影响力，并削弱美国人工智能巨头的优势。中国虽然也会在口头上提及风险和安全，但几乎看不出他们将其视为需要囤积居奇的独特力量，也没有把它当作地缘政治的王牌或具有生死攸关意义的战略武器。For a while (well, for a few months, but that’s an eternity in A.I. time), I’ve wondered about the sincerity of the Chinese position. Since its companies and models have consistently lagged behind Anthropic and OpenAI, China has obvious incentives to use open source to undermine Silicon Valley’s profits and reach, all while letting Washington worry about the risks that the most cutting-edge models may create. Were China to suddenly surpass us, I’ve speculated, the incentives might change, and it might start trying to lock down and weaponize its best models, treat them as state secrets and sovereign goods.过去一段时间（好吧，也就几个月，但在人工智能时代已是永恒），我一直怀疑中国这种立场的真诚性。由于中国的人工智能企业和模型一直落后于Anthropic和OpenAI，中国显然有动力借助开源模式削弱硅谷公司的利润和影响力，同时让华盛顿去操心最前沿模型可能带来的风险。如果有一天中国突然实现反超，我曾猜测，它的激励机制可能会发生变化，它可能会开始试图封锁其最好的模型并将其武器化，将它们视为国家机密和主权资产。We do not yet have a Chinese model that surpasses our own, but this week brought a model that may come close, Kimi K3 from the Chinese company Moonshot AI. And so far Beijing is treating Kimi K3 like all its prior models: It’s going to be open source for maximum adoption, without any American-style fretting over risks and guardrails and export controls.我们目前还没有看到超越美国的中国模型，但本周出现了一款可能已接近这一水平的新模型——来自中国公司月之暗面的Kimi K3。而迄今为止，北京对待Kimi K3的态度与此前所有模型如出一辙：开源，追求最大程度的采用率，没有像美国那样因担忧各种风险而在安全护栏和出口管制上患得患失。This decision should make us a little more confident that Beijing’s overall approach is sincere. It suggests that the Chinese don’t just reject the utopian and dystopian scenarios common in Silicon Valley, where past some threshold the power of artificial intelligence totally remakes the world. They may not even be especially “risk-pilled,” in the sense of worrying about the kinds of perils that today’s A.I. models might already create — through the power to hack and bring down complex systems, the power to help terrorists create bioweapons, the uncertain risks of rogue A.I. behavior and more.这一决定应该让我们对北京整体路线的真诚性多了几分把握。它表明，中国人不仅拒绝接受硅谷流行的那种乌托邦或反乌托邦叙事——即越过某个临界点，人工智能的力量将彻底重塑世界。他们甚至可能压根就不是特别买那些风险论的账，也就是说，他们并不怎么为当下人工智能模型可能引发的危害而焦虑——无论是攻击复杂系统致其瘫痪、协助恐怖分子研制生物武器、失控人工智能行为的不确定风险，等等。American A.I. companies and the national security state seem to assume that we’re already at a point where you simply can’t let the world use the best A.I. models without controls and guardrails any more than you’d ship atomic bombs or ICBMs to random homes or private companies. Whereas the Chinese are willing to shrug and push the envelope and let America fret about safety and sovereignty, on the assumption that A.I. is a normal technology and the safety concerns will get figured out along the way.美国的人工智能企业和国家安全机构似乎认为，我们已经进入了这样一个阶段：世界再也不能在没有监管和安全护栏的情况下自由使用最先进的人工智能模型，就像不能把原子弹或洲际弹道导弹随意送到普通家庭或私人企业一样。而中国则似乎对此并不十分在意，愿意继续大胆推进，让美国去操心安全和主权的问题，因为他们认为人工智能本质上仍是一项普通技术，其安全问题会在发展过程中逐渐得到解决。Or to return to a more cynical reading, maybe they’re making a double bet. If A.I. creates only modest risks and isn’t actually some kind of strategic superpower, their open-source models are the best way to undermine the American advantage and make the world like China more than us. Alternatively, if we do get some kind A.I. Chernobyl from someone misusing a Chinese model — well, the repercussions and backlash will still probably hurt U.S. companies the most, because the American public is already skeptical of A.I. and primed for data center moratoria.当然，如果换一种更加现实主义的解读，中国也许打的是进退自如的算盘。如果人工智能最终带来的风险有限，并没有真正成为某种具有战略意义的超级武器，那么开源模型就是削弱美国优势、让世界更加亲近中国的最佳方式。又或者，即便真的有人滥用中国的某款模型制造出一场类似“人工智能切尔诺贝利”的灾难，那么其冲击和反噬很可能仍然对美国公司的伤害最大，因为美国公众本来就对人工智能持怀疑态度，而且已经做好了支持限制数据中心建设等监管措施的准备。I’m not sure where this leaves the A.I. safety movement, which is responsible for a lot of that pro-regulatory priming. The cause of A.I. safety has advanced along several lines in the United States, benefiting from the progressive left’s hostility to Silicon Valley, the Pentagon’s anxiety about the power of privately controlled models and the A.I. industry’s internal doomer culture.我不太确定这将把人工智能安全运动带向何方，而正是这一运动推动了美国社会如今越来越支持加强监管。在美国，人工智能安全事业的发展得益于几股力量的共同作用：进步派左翼对硅谷的敌意、五角大楼对私人企业掌握人工智能模型能力的担忧，以及人工智能行业内部盛行的“末日论”文化。But it avails the would-be regulators nothing to slow American progress if China believes it’s playing a completely different game.然而，如果中国相信自己参与的是一场完全不同的博弈，那么无论监管者如何努力放慢美国的发展速度，都将无济于事。Ross Douthat自2009年起一直在时报担任观点版面的专栏作者。他也是观点播客“Interesting Times”的主播。他著有多本书，最近的一本是《The Deep Places: A Memoir of Illness and Discovery》。欢迎在Facebook和X上关注他。翻译：纽约时报中文网点击查看本文英文版。