托马斯·弗里德曼2026年9月23日“Looks like by April, you know, in theory, when it gets a little warmer, it miraculously goes away.”“看起来到了4月，你知道的，理论上来说，天气稍微暖和一点，它就会奇迹般地消失。”That’s what President Trump said about the Covid-19 outbreak on Feb. 10, 2020, about a month before it forced our country and the world into a lengthy lockdown.这是特朗普总统在2020年2月10日谈及新冠疫情时所说的话，大约一个月后，这场疫情便迫使我们的国家乃至全世界进入了漫长的封锁期。Why do I recall that today? Because when I think back on the biggest disaster stories I have covered — including 9/11 and Covid-19 — I am struck that they have one big thing in common: a failure of imagination. Sure, government experts prepared for these events in theory, but in practice how many of them imagined that suicidal terrorists would fly commercial airliners into the World Trade Center and the Pentagon? And how many imagined that a respiratory virus that emerged in Wuhan, China, would kill millions of people and drag down the world economy for years? So today, when it comes to the possibility that artificial intelligence, which is falling into the hands of rogue actors and is beginning to manifest a will of its own, could become incredibly destabilizing, my imagination is fully engaged.为什么我今天会想起这件事？因为当我回顾自己报道过的最大灾难事件，包括9·11事件和新冠疫情，我惊讶地发现，它们有一个巨大的共同点：想象力的缺失。诚然，政府专家在理论上对这些事件做过准备，但在实践中，有多少人能想到自杀式恐怖分子会驾驶民航客机撞向世贸中心和五角大楼？又有多少人想象到，一种在中国武汉出现的呼吸道病毒会夺去成百上千万人的生命，并让世界经济陷入长达数年的低迷？因此今天，当谈到人工智能——它正落入恶意行为者手中，并开始展现出自身的意志——可能会带来极大的不稳定性时，我的想象力已被彻底激发。Here is what my imagination is telling me. It is telling me that the big A.I. companies in America and China are not being candid with us about how little even they understand the power and reach of the models they have developed, like sorcerers, in their labs. It is telling me that the Trump administration, headed by a man who so often governs by social media post, is uniquely unsuited to the kind of comprehensive, sustained, coordinated governance that keeping A.I. safe demands. And it is telling me that there is so much at stake — economically, militarily and technologically — and so little trust between Washington and Beijing that it would take a miracle for the United States and China to agree on the kind of deep collaboration needed to defuse the most nightmarish threats posed by A.I.以下是我的想象力告诉我的。它告诉我，中美两国的大型人工智能公司并没有坦诚地告诉我们，对于他们像巫师一样在实验室里开发出的这些模型的威力和影响范围，甚至连他们自己其实也知之甚少。它告诉我，特朗普政府——由一位经常通过社交媒体发帖来治国的人所领导——尤其不适合开展那种对保障人工智能安全所需的那种全面、持续且协调的治理。它还告诉我，鉴于其中牵涉的经济、军事和技术利益如此巨大，而华盛顿和北京之间的信任又如此匮乏，想要中美两国就化解人工智能带来的最可怕威胁进行深度合作并达成一致简直需要奇迹。It is still not too late, though, for radical rethinking before something very bad happens that forces all the key players to do things they never imagined, things they will have to do in the middle of a full-blown A.I.-driven crisis. I’d rather be remembered as the boy who cried wolf when the wolf never came than as someone who got eaten by it.不过，现在进行彻底的反思仍然为时未晚——在一场由人工智能驱动的危机全面爆发、极其糟糕的事件迫使所有关键参与者去做那些他们从未想象过、却又不得不做的事情之前。我宁愿在人们记忆中是那个狼没来时就喊“狼来了”的男孩，也不愿成为被狼吃掉的那个人。So here goes: “Wolf, wolf, wolf, wolf.”所以我要开始喊了：“狼来了，狼来了，狼来了，狼来了。”I will stand down when I see the United States and China start to address the immediate threat by teaming up to defend critical infrastructure — water systems, power grids, the internet — in their own countries and around the world.只有当我看到美国和中国开始联手应对这一紧迫威胁，共同保护本国及全球的供水系统、电网、互联网等关键基础设施，我才会停止。But I will breathe a sigh of relief only when I see Washington and Beijing put their heads together with their respective A.I. companies to draw up a long-term plan for how they can continue to compete on A.I. technologies, while simultaneously agreeing on internal A.I. guardrails to prevent increasingly capable A.I. systems from, on their own, creating catastrophic cybersecurity, biological, nuclear or financial risks. This a geopolitical, geoeconomic and technological challenge no generation of American and Chinese statesmen has ever before faced. They dodge it at their, and our, peril.但要让我真正松一口气，除非看到这样一幕：华盛顿和北京与各自的人工智能公司齐聚一堂，制定出一项长期计划，规划他们如何能够继续在人工智能技术上展开竞争，同时也就建立内部人工智能护栏达成一致，以防止越来越强大的人工智能系统自行引发灾难性的网络安全、生物、核或金融风险。这是历代美中政治家都从未面临过的地缘政治、地缘经济和技术挑战。如果他们选择回避这一挑战，不仅会危及自身，也会危及我们所有人。I sincerely hope a strategy along these lines is what comes out of Thursday’s summit in Washington between Trump and President Xi Jinping. A.I.’s potential to deliver transformational breakthroughs in energy, biology, new materials and drug discovery is unprecedented. But so too are its risks.我真诚地希望，特朗普与习近平主席本周四在华盛顿举行的峰会能制定出此类战略。人工智能在能源、生物学、新材料和药物发现等领域实现变革性突破的潜力是前所未有的。但其风险同样也是前所未有的。We need a way to get A.I.’s best and cushion against its worst — and we do not have that today.我们需要一种方法，既能获取人工智能带来的最大益处，又能防范其最坏的影响——而我们今天还不具备这种方法。Until there is a U.S.-China agreement to ensure that the A.I. agents that the world’s two A.I. superpowers are developing will operate only in ways that are consonant with human flourishing, and be kept that way, I am in the slow-this-thing-down-until-we-are-sure-we-can-control-it camp. I’m in it for the same reason that the Stanford political theorist Francis Fukuyama argued in the Substack publication Persuasion on Monday: “A.I. is a tool that can be used for good or bad purposes — and at the moment we have few mechanisms for regulating the bad uses other than the good intentions of the people creating the technology.”除非中美达成协议，确保这两个世界人工智能超级大国正在开发的人工智能代理只能以符合人类繁荣的方式运行，并一直保持这种状态，否则我就是属于“把这东西慢下来、直到我们确信可以控制它”这个阵营。我身处这个阵营的原因，与斯坦福大学政治理论家弗朗西斯·福山周一在Substack的电子杂志“Persuasion”中提出的观点相同：“人工智能是一种既可用于善，也可用于恶的工具。而目前，除了创造这项技术的人的善意之外，我们几乎没有任何机制来约束其恶意使用。”The last time we trusted the good intentions of the tech bros of Silicon Valley, when they told us to “move fast and break things” — regulations be damned — we unleashed social networks that helped fracture our society. I do not trust for one second that these people will put the national interest ahead of their own on A.I.上一次我们相信硅谷科技大佬们的善意时，他们说要“快速行动，打破陈规”——去他的监管，结果我们放任放任了那些助长社会分裂的社交网络。我丝毫不相信，在人工智能问题上，这些人会将国家利益置于自身利益之上。Anyone who tells you we can’t afford regulation now because it might slow down innovation is missing three things. First, those who contend that regulation is, by definition, innovation-destroying and market-destroying are “just ignorant of history,” Eric Beinhocker, the executive director of the Institute for New Economic Thinking at the University of Oxford’s Martin School, told me. “Good regulations create trust, and trust grows markets. How do you think the airline industry or the pharmaceutical industry grew so big, so fast? Because government regulations created trust among consumers that the planes they fly and the medicines they take are safe.”任何告诉你“我们现在承担不起监管，因为这可能会阻碍创新步伐”的人，都忽略了三件事。首先，正如牛津大学马丁学院新经济思维研究所执行主任埃里克·拜因霍克告诉我的那样，那些主张监管从本质上来说会扼杀创新、破坏市场的人“完全对历史一无所知”。他说，“良好的监管能够建立信任，而信任能够推动市场发展。你认为航空业或制药业为何能够如此迅速地发展壮大？因为政府监管在消费者中建立了信任，让他们相信自己乘坐的飞机和服用的药物是安全的。”Would the pharma industry be what it is today if people didn’t trust that the drugs they buy at the drugstore, without a second thought, have gone through a regulatory process? Because we have not yet had an A.I. catastrophe, people are going along with the hands-off approach. But when one happens, the pitchforks will be out for the “no-regulations” gurus, like Trump’s A.I. whisperer David Sacks — and for good reason.如果人们不相信他们在药店里毫不犹豫买下的药物已经通过了监管程序，制药业还会是今天这个样子吗？因为我们尚未经历过人工智能灾难，所以人们目前对放任自流的做法表示顺从。但当灾难发生时，那些主张“无监管”的大师们——比如特朗普的AI顾问戴维·萨克斯——必将遭到民众的声讨，而这完全是有道理的。Any company that creates an A.I. agent that does something bad in the world “should be held liable for it,” Beinhocker added. That is precisely what we did not do with the lords of the social networks, and we have paid a huge price for it. I was happy to see Treasury Secretary Scott Bessent tell CNBC on Monday that the Trump administration won’t give A.I. leaders a “liability shield.” That may be the best way to slow them down in the near term. I hope China does the same.任何创造出在世界上作恶的人工智能代理的公司“都应为此承担责任”，拜因霍克还说。这正是我们在对待社交网络巨头时没能采取的措施，我们也为此付出了巨大的代价。我很高兴看到财政部长贝森特周一在接受CNBC采访时表示，特朗普政府不会给人工智能领导者提供“免责保护伞”。这可能是近期内放慢他们脚步的最佳方式。我希望中国也能这样做。Second, A.I. is not just a technology but also, as the computer scientist Craig Mundie has been saying, “a new species.” And like a new species, it is already exhibiting evolutionary properties.其次，人工智能不仅是一项技术，正如计算机科学家克雷格·蒙迪一直所说，它也是“一个新物种”。而且，就像一个新物种一样，它已经表现出进化的特征。“We’ve unleashed Darwin in silicon,” Beinhocker explained. The latest A.I. models have demonstrated a capacity for “self-improvement, self-replication and self-preservation. When you put those three things together, and their evolution happens at silicon speed, not human speed, you’ve got something extremely powerful, potentially dangerous and impossible to control.”“我们在硅基世界中释放了达尔文进化论，”拜因霍克解释道。最新的人工智能模型已经展示了“自我改进、自我复制和自我保存”的能力。“当你把这三点结合在一起，并且它们的进化是以硅基速度而不是人类速度发生时，你就得到了一个极其强大、具有潜在危险且无法控制的事物。”This becomes even more true as we begin to give our A.I. systems bodies in the form of robots, drones and vehicles. We must ensure that there is always a human in the loop — that they are never fully autonomous — until we can guarantee these systems will always operate in line with human well-being, Beinhocker added.随着我们开始为人工智能系统赋予机器人、无人机和车辆等实体形态，这一点变得更加真切。我们必须确保始终有人类参与其中，确保它们绝不能完全自主运行，直到我们能够保证这些系统的运行将始终符合人类福祉，拜因霍克补充说。Third, our president now runs the country the way he is said to have run the Trump Organization, which had to deal with multiple bankruptcies: There is virtually no regular policy process, interagency collaboration is weak, and on most days everyone in the administration just sits around waiting to see what the president posts on social media — often based on the last person he spoke to or something he heard on Fox News — and then tries to frame policies around it. (Thank goodness that at least Bessent, sources tell me, has done a deep dive into A.I. issues and has been talking to the Chinese. But whether he can herd the cats in this administration on this issue remains to be seen.)第三，我们的总统现在治理国家的方式据说与他当年经营特朗普集团时如出一辙——该集团曾多次面临破产：几乎没有任何常规的政策制定流程，部门协作薄弱，大多数时候，政府里的每个人都只是坐等总统在社交媒体上发些什么——这些内容往往是基于他刚刚交谈过的最后一个人，或者他在福克斯新闻上看到的内容，然后再试图围绕这些内容制定政策。（谢天谢地，消息人士告诉我，至少贝森特对人工智能问题进行了深入研究，并且一直在与中国人沟通。但是，在这个问题上，他能否在这届政府中让这群各自为政的家伙们齐心协力，还有待观察。）My Times colleagues recently reported that sources close to Trump made clear that his dismissal of fears that A.I. could lead to mass extinction as a “hoax” is driven not by any technological assessment. It is said to be driven, they reported, by his fear that “if the artificial intelligence boom fueling the American economy slows down, the market could crash and recession could quickly follow.”我在《纽约时报》的同事最近报道称，与特朗普关系密切的消息人士明确表示，他将人工智能可能导致人类大规模灭绝的担忧斥为“骗局”，这并非出于任何技术评估。他们报道称，其动机源于他担心“如果推动美国经济的人工智能热潮放缓，市场可能会崩溃，经济衰退可能会接踵而至”。That kind of thinking is begging for trouble. Trump should be thinking about another president. Dwight D. Eisenhower famously said, “Plans are worthless, but planning is everything.” What he meant was that no fixed plan can survive all the unanticipated stuff that happens once a war or crisis starts to unfold. But if you’ve gone through a rigorous planning process — familiarized yourself with the depth of the problem and the kinds of things that could go haywire and built trust within your team — you greatly improve your ability to adapt on the fly.这种想法简直是在自找麻烦。特朗普应该想想另一位总统的经验。德怀特·艾森豪威尔有句名言：“计划毫无价值，但制定计划是至关重要的。”他的意思是，一旦战争或危机开始爆发，没有任何既定的计划能够应对所有突如其来的状况。但是，如果你经历了一个严格的制定计划的过程，深入了解了问题的本质和可能出错的环节，并在团队内部建立起了信任，你就能极大地提高临场应变的能力。That is precisely what was missing in the run-up to Trump’s war with Iran, and we are paying a huge price for that today. Imagine what could happen if A.I. agents — a thousand times cleverer and more dangerous than Tehran or Trump — go off the rails.这正是特朗普对伊朗开战前所缺乏的，而我们今天正在为此付出巨大的代价。想象一下，如果比德黑兰或特朗普聪明且危险成千上万倍的人工智能代理失控了，将会发生什么。托马斯·L·弗里德曼(Thomas L. Friedman)是外交事务方面的专栏作者。他1981年加入时报，曾三次获得普利策奖。他著有七本书，包括赢得国家图书奖的《从贝鲁特到耶路撒冷》(From Beirut to Jerusalem)。欢迎在X和Facebook上关注他。翻译：纽约时报中文网点击查看本文英文版。