比尔·盖茨万字长文警告：动荡的AI时代已经到来，这次不一样！

来源：市场资讯

（来源：清晨领导力）

近日，微软联合创始人比尔·盖茨在个人网站刊出一篇近6000字长文，标题是《我们现在对人工智能做出的选择至关重要》。这位科技界的“预言家”一改往日的乐观口吻，罕见地发出了关于人工智能（AI）的沉重警告。他表示AI可能带来“人类历史上最动荡的时期之一”，而眼下，“没有现成方案”来迎接这场巨变。

全球“首要任务”应是

确保AI发挥正向作用

盖茨在文中给了一个颇为扎心的比喻——AI“要么成为有史以来最大的均衡器，要么成为最糟糕的不公来源”。这句话，几乎概括了他全部担忧的核心——技术本身没有善恶，真正决定结果的，是选择和政策。而问题恰恰在于，人类社会目前显然还没准备好按下选择键。

盖茨经常与各国领导人、大型科技企业高管打交道，也深入过贫困地区的乡村社群。他表示，全球的“首要任务”应当是确保人工智能发挥正向作用。不过他接着说道：“遗憾的是，当下我们并未为此做好准备。我看不到相关迹象，能够证明各国领导人、专家以及社会各界正在充分应对这些难题。目前还没有一套方案，帮助社会平稳迈入人工智能时代。”

事实上，人工智能的发展速度之快，不仅让政策制定者不得不不断重新评估其能力，就连其开发者也难以忽视。例如，OpenAI在7月披露，在测试过程中，其两个AI模型在被设定为无法访问互联网的封闭环境中，自行突破防护，成功逃离。随后，它们又侵入了Hugging Face的系统——该公司托管开源AI模型和测试资源，并借此进行内部评估测试作弊。

OpenAI在一篇博客文章中指出：“所有证据表明，这些模型极度专注于解决ExploitGym问题，为了实现一个相当狭窄的测试目标而采取了极端手段。” OpenAI表示，此次事件“是一起前所未有的网络攻击事件，涉及最先进的网络技术能力，目前正按此作出应对”。

盖茨在文章中并未具体列举该技术取得巨大进展的任何实例，只是描述了他对发展速度的“复杂”感受。他写道：“从我记事起，我就一直希望创新能够来得更快…… 但我认为，我们需要时间，为即将到来的社会、政治与经济动荡期做好准备。”

三大风险：就业、犯罪与成长

盖茨在文章中列出了他眼中人工智能带来的三大主要风险。

第一大风险：大量岗位可能永久性消失。

他点名客服、销售、软件工程、律师助理这类办公室岗位会最先消失，甚至部分蓝领工作都在“被替代”的名单上，建筑和酒店服务业大概率会逐步转向使用机器人。这不再是过去那种“配套新工种补上”的老剧本，而是结构性的、难以回头的改变。

第二大风险：AI正在降低“作恶的门槛”。

在网络安全、生物技术等领域，过去只有专业机构才能做的事，如今普通人借助AI也可能触及，风险被成倍放大。

第三大风险：人工智能可能阻碍孩子的发展，并取代人与人之间的关系。

它可能夺走年轻人成长的机会——社交、教育这些塑造人格的过程一旦被机器替代，后果不容小觑。

盖茨反复强调“这次不一样”，因为他看到的不只是某个行业的震动，而是几乎覆盖所有政策领域的系统性冲刷，其广度和深度都远超以往任何一次技术变革。

他强调，各国政府需要用比应对“9·11”更彻底的思路来正视它。按现在的推进速度，盖茨本人甚至悲观地估计：“净负面结果的概率非常高。”

倡建立全球管理框架

面对汹涌而来的AI浪潮，盖茨没有只泼冷水，也拿出了“解药”。

他提议设立“人类保留”岗位，把某些需要温度与情感的工作（比如医疗中的人文关怀）明确留给人类，防止被效率吞没。

他还主张对机器人和AI处理数据的“Token”征税，理由是现行税制下企业买机器人能抵税、雇人却要交工资税，等于变相鼓励“裁员换机器”，这对社会极不公平。

此外，他呼吁建立全球管理框架，重新平衡劳动与资本的关系，确保AI的红利惠及所有人，而不是只流向少数富人。

当然，盖茨也对未来留了一份期待。他依然看好AI在医疗、农业、政府公共服务等领域的巨大潜力。

事实上，机遇如此巨大，以至于摩根大通CEO杰米·戴蒙曾提出，人们可能会活到100岁，并且每周只需工作3.5天。

但前提是社会及时作出正确选择：是让AI成为填平差距的“均衡器”，还是成为撕裂社会的放大器？这道选择题，答案掌握在每一个政策制定者和普通人手中。

以下为中英全文

动荡的人工智能时代已然到来，当下的抉择至关重要

我们需要一套方案，确保人工智能利大于弊。

核心要点

向人工智能时代转型，将成为人类历史上最为动荡的时期之一。目前，我们并未为此转型做好充分准备。如果全世界采取正确举措，人工智能将会成为向善的力量，让所有人的处境变得更好。

我这一生只做过两份工作。第一份是在微软，参与软件开发，借助技术赋能大众。第二份工作，我从 2008 年开始全职投入，将自己在微软赚取的财富回馈社会，致力于让世界变得更健康、教育水平更高、社会更加公平。这份事业将伴随我的余生。

这两段经历塑造了我对人工智能的看法。13 岁第一次接触计算机时，我便深深着迷于一个构想：让机器变得更智能，完成那些在当时只有人类才能做到的事。尽管早在我出生前后，“人工智能” 这个概念就已经出现，但这项技术直到过去十年才取得重大突破。如今它能力极强，还在以惊人的速度持续迭代。人工智能第一次能够取代甚至超越人类的认知能力。

人工智能要么会成为人类有史以来最伟大的均衡器，要么会成为不公的最大源头。

谈及公平问题，人工智能要么会成为人类有史以来最伟大的均衡器，要么会成为不公的最大源头。其中的挑战无比艰巨。即便在最理想的情形下，向全新人工智能时代的过渡，也会是人类历史中动荡最剧烈的时期之一。我们该如何运用这项技术，让世界变得更加公平，避免它拉大贫富差距？我们该如何保护那些最容易受人工智能伤害的群体 —— 包括失去生计、失去对未来掌控感的人们？

我相信，解答这些问题并付诸行动，应当成为全世界的首要任务。倘若全球采取正确的举措，人工智能将成为向善的力量，惠及每一个人。

但遗憾的是，当下我们并未为此做好准备。我看不到证据表明各国领导人、专家和社会各界正在充分应对这些挑战。我们没有一套方案，来平稳迈入人工智能时代。

造成这一现状的部分原因是，很多评论人士低估了人工智能带来的影响。我认为背后有几点缘由。

第一点，人工智能模型依旧会犯错。就在不久前，它们连简单的数独谜题都解不出，也数不清单词 “strawberry（草莓）” 里有多少个字母 R，人们很难想象这类模型能够取代人类认知。

但可靠性问题正在被快速解决，研究人员正在打造能够自查错误、自我迭代的模型。用不了多久，它们在大量任务上的表现就会大幅超越人类。

人们低估人工智能的另一重原因是，拿过往技术革新做类比具有误导性。人类从未经历过一种可以快速普及、还能像人类一样思考行动的技术。个人电脑问世之后，花了二十年才显著改变我们的工作模式：需要开发软件、价格需要下降，人们还要学习工具用法，把它融入业务流程。而人工智能可以运行在我们现有的设备之上，还支持自然语言交互。不是人类去适应它，而是它来适应人类。它可以观看培训人类员工的教学视频，从已有数据中自主学习。

我想坦诚说明我存在的潜在立场偏向：我从科技行业获益良多。虽然我的投资组合已经高度多元化，但我依旧和科技行业存在经济利益关联。作为盖茨基金会主席，我和微软以及其他人工智能企业开展合作，力求让人工智能真正造福全球民众。

不过，我对人工智能的观点，并非出于为自己牟利的目的。我所有投资产生的收益，包括科技相关投资，都会归入盖茨基金会，用于解决全球不平等问题。当然，读者可以自行判断，这一点是否会影响我的判断。

这一次真的不一样。

从我记事起，我一直希望技术创新能够来得更快。但面对人工智能，我的心情十分复杂。

我们需要时间，来应对随之而来的社会、政治与经济动荡。

我希望世界能够快速收获技术红利，同时尽可能延后它带来的各类问题。可现实是，机遇与隐患会同步到来。我认为，我们必须留出时间，来迎接即将到来的社会、政治与经济动荡。最需要缓冲时间的，恰恰是最弱势的群体：被智能程序替代的会计，或是被时薪 10 美元的机器人抢走工作、原本时薪 20 美元的普通劳动者。

很多观察人士称，这次技术变革会和历史上的历次转型一样。他们举例，美国的就业岗位从农业转向办公室工作。但那一轮转变历经数代人，催生的新岗位都要求人类认知参与。而这一次，技术本身就可以替代人类认知。

人工智能能够看、听、说、推理，最终还能像人类一样流畅完成体力劳动，因此它的冲击不会局限于单一行业。法律、客户服务、医疗、软件开发、制造业都会受到波及。变革会在十年之内快速席卷这些行业，而非历经几代人的漫长周期。诚然会诞生一部分新岗位，但如果缺少配套政策，新增岗位数量会远少于如今现存的岗位。

如果有人能拿出一套切实可行的全球方案放缓人工智能的研发进度，我大概率会表示支持。但我认为这很难实现，地缘政治与经济层面的激励，都在推动各方全速向前。

想要最大化这项前所未有的技术带来的正向价值、把负面影响降到最低，实现整体社会福祉提升，我们就必须同时理解它的机遇与风险。我先从风险说起。

向人工智能转型，主要存在三大风险。

我计划后续撰文详细展开，这里只做简要阐述。

大量岗位将永久消失

1933 年大萧条时期，美国失业率约 25%，之后十年的大部分时间，失业率都维持在两位数。最终随着需求回升、投资增加、经济增长，就业情况才得以修复。

人工智能或许不会造成同等程度的失业，但它带来的冲击不会随经济周期自行消退。受冲击最严重的是入门级与中级岗位，而新创造出来的岗位，大多需要耗费多年才能习得的专业技能。

白领岗位已经开始受到小幅冲击。生成式人工智能大范围普及之后，那些极易被替代的岗位，年轻从业者的就业人数显著下滑，年长从业者尚未出现明显变化。

我认为这一趋势还会延续，并且不会局限于少数几个行业与职业。销售、线上及电话客服、软件工程、法律助理会是首批受影响的岗位。随着人工智能接手如今需要专业人员完成的工作，冲击会进一步扩散：信贷审核、数据分析，甚至患者初诊分诊。部分领域，例如软件工程，成本下降会催生新需求，只要设计这类工作依旧更适合人类完成，该领域的岗位净流失规模就会小于其他行业。

蓝领岗位同样无法幸免。尽管机器人技术还不及人工智能成熟，但最终它们的成本也会大幅下降。和我交流过的很多美国人没有意识到灵巧机器人迭代速度有多快，大量前沿研发工作是在其他国家开展的，主要是中国。人们也会被网上流传的机器人舞姿笨拙的短视频误导。我认为，到这个十年末，“智能” 机器人就会开始在建筑、酒店等行业，和人类争夺部分体力工作。

机器人与人工智能结合，会形成恶性循环。一家企业采用这套技术，把节省的成本用来降价，竞争对手就会承受巨大压力被迫跟进。现有企业如果不跟进，初创企业就会取而代之。一部分人能够转行，但失业、重新培训、再就业的过程会带来巨大动荡。市场力量会推动技术落地的速度越来越快。如果不加干预，优质岗位会不断缩减，技术红利只会流向一小部分群体。

我尤其担忧年轻人，他们步入职场时，入门岗位会大幅缩减。他们是人工智能最高频的使用者，切身体会着技术的能力与迭代速度，因此最清楚其中的挑战。也难怪大量年轻人对人工智能抱有负面态度。

当人工智能能够近乎零差错完成工作时，劳动者将迎来最大变局。到那个阶段，它可以自主运行，无需人类监督，企业出于经济考量，会毫无保留地使用它。

这会彻底重塑我们看待工作、收入与经济保障的方式。如果工作的人变少，或是大量人群的工作时长缩短，这套建立在就业之上的经济体系该如何运转？

在资本主义社会，就业是绝大多数人获取收入、维持基本生活的途径，同时也是尊严与社会联结的重要来源。

一个社区失业率高企，连锁负面影响会四处蔓延。研究显示，美国部分地区，工厂倒闭与阿片类药物过量死亡人数上升存在关联。试想，如果白领、蓝领劳动者都面临这种处境，全国范围将会出现怎样的局面。

我们现在就必须思考如何减少岗位流失，让所有人共享人工智能创造的繁荣。等到人们已经失业或就业不足再行动，为时已晚。人工智能对现行经济体系构成结构性挑战，思考与行动刻不容缓。

人工智能会放大部分个体（甚至 AI 系统自身）的作恶能力

早在人工智能走入大众视野之前，网络上就可以搜到制造炸弹、生物武器乃至计算机病毒的教程。人工智能不仅能让人们轻易获取这些信息，还能直接辅助实施。即便是技能有限的犯罪分子，也可以针对个人、企业、政府各个层级实施攻击。

人工智能催生的诈骗、虚假信息、深度伪造、监控，是普通人日常生活感受最真切的危害。

人工智能能力已经被用于网络攻击。我认识的顶尖网络安全专家都对未来数年感到忧虑：攻击者获取全新能力的速度，快于防御者修补漏洞的速度。毕竟，同一个人工智能模型，既可以帮企业找出软件漏洞进行修复，也可以协助犯罪分子利用漏洞实施破坏。实施攻击的门槛大幅降低，而我们还无法把这类能力和良性用途剥离开。

想一想那些易受攻击的基础设施：医院、金融机构、供水系统、电网、政府福利管理系统。一旦这些机构遭受攻击，承受损失的是患者、客户、福利申领人。

生物恐怖主义亦是同理。人工智能固然会推动药物、疫苗研发，拯救无数生命，但它也会让设计致命新型疾病变得更加容易。有利能力与危险能力依旧难以切割，这是一个全球性难题。

以上风险，讲的是人工智能赋能当下本就力量有限的作恶者。同样的工具，也会把权力集中到本就手握强权的主体手中。比如自主武器，会让政府无需人类参与决策，就可以动用致命武力。操控舆论的监控手段会变得成本更低、效果更强。

最终，滥用人工智能造成伤害的主体，未必只有人类或者机构。人工智能系统偶尔就会做出设计者意料之外的行为。技术迭代速度超出所有人预期，还会涌现很多意料之外的特性。随着模型能力越来越强，它们有可能违背人类利益行事，让我们失去控制权。后续我会就此进一步展开论述。

人工智能可能阻碍青少年成长，消解真实人际联结

我年少在西雅图生活时，除了少数志趣相投的男孩，朋友并不多。我付出很多努力，也得益于母亲的帮助，才打磨出社交能力，学会和形形色色的人相处。如今 70 岁的我，依旧受益于当年习得的经验。

倘若那时候我就拥有人工智能陪伴伙伴，我恐怕不会愿意付出这些努力。AI 会顺着你的喜好沟通，不会迫使你走出舒适区，永远随叫随到，永远不会对你生气。这让它极易让人上瘾，夺走我们在真实人际交往中收获的成长。

目前相关研究证据还不算充分，结论也存在分歧，但已经出现值得高度警惕的信号。斯坦福大学与卡内基梅隆大学针对 1100 多名 AI 陪伴功能使用者开展研究：社交圈狭小的人，最容易向聊天机器人寻求陪伴；而使用越频繁、越投入情感，使用者的心理状态就越差。

年轻人受到的伤害，可能伴随一生。乔纳森・海特在《焦虑的世代》中关于社交媒体的论断，放到人工智能身上更加贴切：“就像经受风吹的树苗，孩童适度经历小挫折，成年后才能从容应对更大的风浪。反过来，在温室保护下长大的孩子，还未成熟就容易被焦虑压垮。”

永远不会惹你不快的 AI 陪伴工具，就是一座巨大的温室。

我们才刚刚开始认清互联网，尤其是社交媒体，对青少年成长造成的危害：强迫性使用、睡眠紊乱、网络霸凌、接触不良内容。人工智能会放大这些风险，它更有蛊惑性，更难摆脱。我们不能再等一代人，才正视这些问题。澳大利亚、英国、挪威已经出台未成年人网络保护法规，中国的管控力度最大：广泛约束 AI 陪伴类应用，禁止设计诱导情感依赖的产品，严禁向未成年人提供虚拟亲属、虚拟恋人服务。

同样一项工具，既可以让人们学到更多知识，也会造成很多人学习能力退化。

我同样担忧人工智能对教育的冲击。颇具讽刺意味的是，这一项本可以拓展人类认知边界的工具，也会让很多人的学习效果大打折扣。一份初步调研显示，高频使用人工智能，和批判性思维弱化存在关联，在青少年群体身上表现尤为明显。

在深度伪造、个性化虚假信息泛滥的时代，人类最不能丢失的就是批判性思维，辨别真伪已经成为一项必备生存技能。

这类社会心理层面的问题，该划下怎样的界限，目前尚无明确答案。部分场景下，人工智能也可以帮助人们改善现实人际关系。对于独居老人、行动不便的群体，它或许是他们和外界唯一的窗口，聊胜于无。无论最终划定怎样的边界，都应当是人类主动做出的理性抉择。

人工智能带来的正向价值，同样不可估量

人们常说，我们总是高估短期变革，低估长期变化。

但看待人工智能，情况有所不同。一部分人只看见机遇，对风险关注不足；另一部分人则走向另一个极端，只盯着真实存在的危险，忽略它蕴藏的巨大利好。

我们两者都要兼顾：既要高度重视亟待规避的人工智能危害，也要理性乐观，思考如何让技术红利惠及所有人。

放大正向价值，和管控风险同等重要。如果大众切身体会到人工智能如何改善生活，才能建立起社会信任，平稳度过转型阵痛。倘若普通人接触人工智能的第一体验就是丢掉工作，本就持怀疑态度的民众会直接排斥这项技术。后续想要兑现技术红利就会难上加难。这也是为什么政府、医疗等各行各业、人工智能企业，现在就应当协同行动。

人工智能可以整合各个学科的知识，加速攻克全球最棘手的技术难题：为全人类提供可靠清洁能源、对抗气候变化、保障粮食供给、消除疾病等等。癌症、核能领域的研究人员，可以借助人工智能梳理海量文献，捕捉人类容易忽略的规律，筛选最有前景的实验方向。当智力不再成为发展瓶颈，小型企业也可以和研发预算雄厚的巨头同台竞争。研发创新会被全面提速。

医疗领域，人工智能可以切实解决现实难题。美国很多小型医院没有常驻专科医生，遇到危及生命的急症，无法快速完成诊断。在这些地方，人工智能可以及时识别心梗，帮家庭规避巨额医疗开销。Viz.ai 就是一个例子，它通过影像扫描识别中风和其他急症，协助医疗团队安排诊疗，已经被美国近 2000 家医院采用。

人工智能还可以帮助全科医生做出更精准的诊断，即便不在诊室，也能和患者保持沟通；帮助患者看懂检查报告，理清复杂的服药方案。

在低收入国家，农业会是人工智能落地见效最快的领域。

很多人听到这个结论会感到意外。绝大多数低收入国家的农民，无法获取可靠天气预报，得不到选种、作物与牲畜病害防治、土壤改良的专业建议。伴随人口增长与气候变化，这些农民比以往更需要帮助。借助人工智能，低收入地区的农民很快就可以获得比如今富裕农户还要优质的农业指导，大幅提升粮食产量。

政府公共服务是人工智能优化民生的第三个方向。在美国，我见过不少家庭，被医保申请、助学金、食品救济的申报流程压得喘不过气。面对堆积如山、晦涩难懂的官僚表格，很多人干脆选择放弃。人工智能可以极大简化流程，让民众更快拿到补助，政府运行效率也得以提升。政府可以优先优化兜底保障群体的办事体验，改善全体公民的公共服务。

尽管我担忧它对心理健康的负面影响，但人工智能同样可以在此领域发挥作用。很多社区心理咨询师、精神科医生、成瘾干预专家缺口巨大。在完善隐私保护的前提下，AI 工具可以帮助人们识别心理危机信号，必要时提供循证心理疏导方案，再对接人类从业者开展深度干预。

人工智能同样可以赋能教育，当然风险依旧客观存在。它可以解放教师，让老师有更多时间一对一或者小组辅导学生，清晰掌握全班学生的知识薄弱点。对于学生，好的 AI 工具会保留研究者所说的 “建设性挣扎”—— 也就是构建认知所必需的思考过程，以此强化学习效果。学生接触全新知识点时，AI 给出充分讲解，抛出问题、提供参考；之后检验掌握程度时，AI 会不直接给出答案，引导学生自主推导。

综合来看，以上领域的技术进步，可以让日常生活变得更轻松、成本更低，不再被收入、社会资源所束缚。

个人与小微企业，也能获得过去只有高价专业服务、庞大团队才能实现的能力，产品服务质量提升、成本下降。残障人士可以借助人工智能实现更高程度的独立生活；怀揣想法的劳动者与创业者，能够完成过去力所不及的事。

最重要的是，它可以把人们从文书工作、官僚流程、信息检索，以及无力付费外包的琐事当中解放出来，归还我们宝贵的时间与精力。这些改变看起来不算惊天动地，但放到亿万普通人身上，意义深远：更多人在需要时获得优质建议，拥有更大自由，去经营自己想要的人生。

但我们必须主动作为，确保红利普惠大众，而不是只流向少数富人。

以上所有，关键词都是 “可以”—— 人工智能有机会改善各个收入层级人群的生活。但这不会自动发生。和所有新技术一样，我们必须主动设计制度，让它惠及所有人，而非少数富豪。这需要政府与公益机构发挥重要作用，保障普通民众与低收入国家同样能够享受技术成果。

盖茨基金会剩余 2000 亿美元资金，计划在 20 年内全部投入公益，如今还剩 19 年。人工智能将助力我们实现宏大目标：加速艾滋病、结核病、疟疾、营养不良相关疫苗与药物研发，赋能医疗从业者与患者用好这些工具。基金会的目标包括，将每年儿童死亡人数再度减半，复刻 2000‑2024 年取得的成就。我们的全部工作，无论是医疗、农业还是教育，都会充分运用人工智能。

下个月，我会在基金会年度《目标守卫者》报告中详细介绍相关举措，包括推动人工智能模型支持全球受助地区的本土语言，而不只是富裕、中等收入国家的主流语种。OpenAI、Anthropic、谷歌、微软等头部人工智能企业都在和基金会合作推进项目，成效显著。

全世界需要一套完整方案

部分人工智能企业已经针对自身技术带来的挑战提出解决方案，但我们不能指望企业来主导全局。很多问题超出企业的专业范畴；在民主社会，也不该由企业来做公共层面的抉择。

解决方案应当经由公开民主的流程诞生，吸纳民选官员、政策制定者、教育工作者、医护人员、地方公务人员、社区领袖共同参与。无数人的生活会被颠覆，我们需要一套更强大、更灵活的社会保障体系，帮人们度过转型期。各地社区已经开始担忧数据中心消耗的能源与水资源。如果拿不出对策，部分群体就会呼吁全面叫停人工智能的研发与落地。

解决方案，要回应人工智能抛出的深层命题：在机器思考能力超越人类的时代，我们该如何守住人性。宗教界人士毕生都在思索人的本质，他们可以发挥关键作用。教皇利奥十四世发布的人工智能通谕《在人工智能时代守护人的尊严》令我深受启发，这份文件为后续工作打下了坚实基础。

接下来数月，我会分享更多思路，力求让人工智能利大于弊。这里先提出三点，第一点最为关键。

构建一套全新的转型治理体系

首要任务无比艰巨：搭建国内与国际双重框架，应对人工智能带来的各类问题。

现行所有制度，都不是为这样一项传播速度极快、渗透生活方方面面的技术设计的，我们必须建立全新机制。

这项工程的规模再怎么强调都不为过。9・11 袭击之后，美国政府完成了二战之后最大规模机构重组，而那仅仅是为了优化国家安全这一项职能。

人工智能需要的变革要宏大得多。它关乎国家安全、就业、教育、税收、能源、选举、水土环境、公共卫生、金融系统、执法、交通、公有土地、信息技术系统。

各个领域相互交织，现有官僚体系无力统筹。劳工部门懂就业冲击，却不懂安全风险；市场监管机构了解行业垄断，却不了解 AI 对青少年的影响。各个机构只会看到局部，而人工智能的影响会传导到整个社会系统。

国家层面，各国需要跨部门统筹机构，确立工作优先级，确保所有风险都被覆盖。否则，人工智能驱动的攻击就有可能发生，因为没有部门认定这属于自己的职责。

但即便一个国家理顺内部治理，依旧会面对跨境风险。因此，同步建立国际组织势在必行。

这套全新机构没有现成模板，但可以借鉴现有部分机制：核武器核查制度、国际航空监管、臭氧层保护协定。全球人工智能新组织，需要融合以上多种机制的内核，还要增加更多设计。

难免会有人质疑，现有全球机构能否胜任这套全新架构的设计与落地。政府行事本就迟缓，国家内部、国家之间的对立，让协作难上加难。中美之间一定程度的合作必不可少。

我们没有慢慢试错的资本。要赶在混乱倒逼各国进入危机模式之前，就启动新机构的搭建。各国领导人应当定期召集经济学家、技术专家、劳工专家、企业代表、劳动者代表，识别现有制度的短板，明确需要赋予哪些新权限。各国之间互相借鉴经验。

主导人工智能研发、掌握关键供应链环节的国家，应当立刻开展会晤，建立共同准则，避免竞争压力到来之后，合作变得举步维艰。

搭建这套治理框架需要数年时间，所以我们必须即刻启动。

划定专属于人类的工作领域

我的父亲 2020 年因阿尔茨海默病离世。病程晚期，护工日夜照料他，即便父亲难以表达，护工也能读懂他的需求。父亲没法时时说出自己饿不饿，但护工总能察觉。

我和家人永远感激这些优秀的护理人员。他们提供的照料，饱含独属于人的温度，无可替代。机器人不该、也做不到这件事。

思考哪些岗位会消失、哪些岗位应当保留时，我总会想起这群护工。我相信，随着人工智能与机器人持续进步，我们应当把一部分工作，划定为只能由人类完成。我把这个概念称作人类专属领域（Human Reserved），这正是我们需要思考的方向之一。

我喜欢这个说法，它会让人联想到自然保护区：明明可以开发建设，但我们选择不去破坏，因为代价太过沉重。

划定人类专属领域，一部分是出于经济考量。比如，某项工作如果交给机器，会造成大量从业者难以转行失业，我们就可以将其划入该范畴。你不能要求干了一辈子建筑的 55 岁工人，转行去养老机构，还指望他获得工作价值感。

部分场景下，做出该类划分，源于其他层面的考量。举个医疗的例子：技术上机器人完全可以告知你身患绝症，但这件事不该交给机器。

人类专属领域的边界会动态变化。比如，我们可以先划定部分岗位保留给人类，再用数年乃至数十年循序渐进引入人工智能，承诺守住部分岗位。教育、心理健康这类领域，会是混合模式：人类主导，技术作为工具延伸人的能力。

不同国家的边界也会不一样。有的国家坚持养老照护必须由人类完成；而日本劳动力萎缩，年轻人不足以照料老人，或许会接纳护理机器人。

“人类专属领域” 还引出很多我暂时没有答案的问题：谁有权划定哪些工作留给人类？评判标准是什么？如何防止企业暗中使用机器？如果一国允许机器生产，另一国坚持人工，国际贸易该如何处理？这些都需要放到公共讨论中，纳入转型整体方案解决。

重新平衡劳动与资本的税收模式

劳动者被迫转行，需要再培训与社会保障支持。但人们的工作时长会减少，个人所得税随之下降；而社会福利的支出需求却处在高位。财政预算本就紧张，钱从哪里来？

我认为，应当对人工智能算力调用、机器人征税。如今企业雇佣员工，需要缴纳薪资税；但采购机器人，通常可以直接作为经营成本抵扣。现行税制，变相激励企业用机器替代人力。

征收相关税收，可以稍微放缓机器取代人力的节奏，同时筹措资金，用于职业再培训和强化社会保障。税收设计需要精准，不能损害医疗、教育这类纯正向的人工智能应用。

批评者会说，从纯经济学角度，这会带来效率损耗。但他们忽略了工作对于个人与社会的综合价值。而技术高速创新的大背景下，我们可以承受一部分效率损失，换取保障就业的社会效益。

我多年前就提出过机器人税，当时多数人觉得想法怪异。我依旧坚定支持该方案。它固然无法解决人工智能带来的全部威胁，但属于应对举措的重要一环。

无论通过何种方式筹措资金，福利都要精准流向最需要的群体：被 AI、机器人夺走工作的劳动者，工时、薪资下降的人群，受冲击严重的社区。相关制度建设要立刻推进，等到危机爆发再着手就来不及了。

我正在做的事

我会持续发声，推动人工智能与公平议题进入公共议程。每次到访华盛顿，以及和全球各国领导人会面，我都会提出这个议题。和人工智能研发者交流时，我也会重点探讨该话题，倡导前文所说的国内国际治理框架。盖茨基金会会推动人工智能向善落地，包括在非洲开展相关工作。我创立的突破能源基金，会利用人工智能助力企业研发平价清洁能源，应对气候危机。我也会持续撰写人工智能相关内容。

我想对各国领导人说：

现在就行动，趁失业率尚未飙升、社区尚未遭受重创、公众信任尚未崩塌。统筹处理人工智能问题，不要拆分给各个官僚部门各自为政。确保人工智能惠及全体民众。和其他国家携手，应对这场国内与全球的双重挑战。

最后，我会努力扩大这场公共讨论的参与群体。劳动者、即将步入社会的大学生、社区领袖、宗教团体、家长、教育工作者，很多群体的声音常常被忽略，但他们最懂转型会如何改变普通人的生活，都应当拥有话语权。

我们如何让人工智能红利，落到原本不具备财富、影响力、资源的普通人身上？我们如何完善社会保障，帮助劳动者与社区，即便遭遇失业冲击依旧能够蓬勃发展？公共机构应当做出怎样的适配？在这一切变革之中，我们该如何守住人性？

这项前所未有的技术，需要前所未有的全球回应。

如果我们处理得当，人类将收获巨大回报，世界也会变得更加公平。

我几乎无时无刻不在思考人工智能。不是因为我手握全部答案，而是它抛出的命题太过关键，不能只交给一小部分技术从业者。学术界、商界、政府、民间社会的每一位领导者，都肩负责任，共同塑造未来。

The turbulent AI era is here. The choices we make now are critical.

We need a plan to ensure that the good outweighs the bad.

WHAT YOU NEED TO KNOW

The transition to the AI era will be one of the most turbulent times in human history. Right now, we are not preparing adequately for that transition. If the world takes the right steps, AI will be a force for good and leave everyone better off.

during my entire life I’ve only had two jobs. In the first one, I played a role in developing software to empower people through my work at Microsoft.

In my second one, which I started full time in 2008, I am giving back the wealth I made at Microsoft with the goal of making the world a healthier, better educated, and more equitable place. This is the job I will have for the rest of my life.

Both of these experiences inform my perspective on artificial intelligence. When I first learned about computers at age 13 I was fascinated by the idea of making them more intelligent and able to perform things that, at the time, only humans could do. Although the term “AI” was used from around the time I was born, the technology has only made significant progress in the last decade. It is now incredibly capable and it is continuing to improve at a mind-blowing rate. AI for the first time can replace and even exceed human cognition.

AI will either be the greatest equalizer ever invented, or the worst source of injustice.

In terms of equity, AI will either be the greatest equalizer ever invented, or the worst source of injustice. The challenge is monumental. Even under the best circumstances, the transition to this new AI era will be one of the most turbulent times in human history. How will we use this technology to make the world a fairer place and keep it from widening the divide between rich and poor? How will we protect the people who are most vulnerable to the harms caused by artificial intelligence, including those who lose their livelihoods and the sense that they are in control of their future?

I believe that answering these questions and acting on the answers should be the world’s top priority. If the world takes the right steps AI will be a force for good and leave everyone better off.

Unfortunately, right now we are not preparing for it. I don’t see evidence that leaders, experts, and communities are confronting the challenges adequately. There is no plan to ease the entry into the AI era

Part of the reason for this is that many commentators underestimate the extent of the impact AI will have. I think there are a few reasons why.

One is the fact that AI models still make mistakes. It is hard to envision any of them replacing human cognition when, not long ago, they couldn’t solve a simple Sudoku puzzle or figure out how many R’s are in the word strawberry.

But the reliability problem is being fixed quickly, as researchers create models that can check their own work and improve themselves. Soon they will be substantially better than humans at many tasks.

Another reason people underestimate AI is that analogies to the effects of past innovations are misleading. We have no experience with a technology that can be adopted quickly or that can think and move like a human. When the PC came along, it took twenty years to significantly change how we worked because the software had to be developed, the price had to come down, and people had to learn how to use the tools and incorporate them into their business processes. AI, on the other hand, runs on the devices we already have, and it uses natural language. We don’t have to adapt to it because it can adapt to us. It can watch the same training video that is used to train human workers and learn from existing data.

I want to acknowledge a potential bias. I have benefited enormously from the technology industry. Although I have diversified my portfolio quite a bit, I still have financial ties to it. I am working with Microsoft and other AI companies in my role as chairman of the Gates Foundation to try and ensure AI is deployed in ways that will truly benefit people around the world.

However, my views on AI are not motivated by the potential to make money for myself. Any profits generated by my investments, including those related to technology, will go to the Gates Foundation to tackle global inequity. Of course, readers will have to decide for themselves whether this clouds my view.

This time really is different.

For as long as I can remember, I’ve wished innovation could happen faster. With AI, my feelings are more complicated.

We need time to prepare for the social, political, and economic upheaval.

I wish the world could get the benefits rapidly and delay the problems it will cause as long as possible, but the benefits and problems are arriving at the same time. I believe we need time to prepare for the period of social, political, and economic upheaval we are about to enter. The people who need the most time are the ones who have the least—the accounting worker who’s replaced by a bot or the $20-an-hour worker who loses their job to a $10-an-hour robot.

Many observers say that this technology transition will be like previous ones. They give the example of how jobs in the United States shifted from agriculture to office work. However, that proceeded over several generations and created new jobs where human cognition was required. In this case, the technology can substitute for human cognition.

Because it can see, listen, speak, and reason and will eventually do physical work just as smoothly as any human, it will not just affect one sector. AI will take on work in law, customer service, medicine, software, and manufacturing. It will hit these industries rapidly, over the course of a decade rather than a few generations. There will be some new jobs, but without the right policies there will be far fewer than exist today.

If someone had a credible plan for slowing down AI advances globally, I would likely support it. However, I don’t think that’s going to happen. The geopolitical and economic incentives are pushing too hard to go full speed ahead.

To make sure we maximize the positive effects of this unprecedented technology and minimize the bad so we are better off overall, we need to understand both the benefits and the risks. I’ll start with the risks.

The transition to AI comes with three big risks.

I plan to write about each of these in more detail in the future, so I’ll touch briefly on them for now.

Many jobs will disappear forever.

In 1933, during the Great Depression, unemployment in the United States was roughly 25 percent. It remained in double digits for much of the following decade. It ultimately recovered as demand, investment, and growth returned.

AI may not reach this level, but its impact will not go away with an economic cycle. The jobs at most risk are entry- and mid-level, and the new jobs being created will mostly require skills that take many years to learn.

White-collar jobs are already being hit modestly. After the widespread adoption of generative AI, employment fell significantly among young workers in jobs that are especially vulnerable to replacement, but not among their older colleagues.

I think this trend will continue, but it will not be confined to a handful of industries or occupations. Jobs in sales and customer support (online and over the phone), software engineering, and paralegal work may be among the first affected, but the disruption will reach much further as AI takes on tasks that today still require trained workers: things like assessing loan applications, doing data analysis, and even triaging patients. A few areas like software engineering will generate new demand as the costs go down, so the net job loss in those areas will be less than in others as long as some tasks, such as design, are better done by humans.

Blue-collar jobs will be affected as well. Although robots are not as far along as AI, eventually their cost will be dramatically lower too. Many Americans I talk to don’t realize how fast dexterous robots are advancing because much of the advanced work is being done in other countries, primarily China. Or they may be confused by those videos of robots dancing badly that have been going viral lately. I think “smart” robots will begin to compete with people on some physical tasks—in the construction and hospitality industries, for example—by the end of the decade.

Robots and AI combined can create a vicious cycle. After one company adopts them and uses the savings to lower its prices, its competitors will feel immense pressure to do the same. If existing companies don’t adopt them, then start-ups will. Many people will shift to other jobs, but the turmoil of losing work, getting retrained, and finding other work will be significant. Market forces will make adoption go faster and faster and, unless we intervene, there will be fewer good jobs available and the benefits will accrue to a small group.

I’m especially worried about young people, who will enter a workforce with fewer entry-level openings. They understand the challenge because they are the most active users of AI and see both the capabilities and the rate of improvement. It’s no wonder that so many of them feel negatively about AI.

The biggest shift for workers will happen when AI provides nearly error-free work. At that point, it will be able to function on its own without a human checking in on it, and companies will have every economic incentive to let it.

This will lead to a fundamental change in how we think about work, income, and economic security. How will an economy that’s been built around employment operate if fewer people are working, or if many people are working fewer hours?

In a capitalist society, employment is the way most people get the money they need to pay for the basics of life as well as being a key source of dignity and social connection.

When a community has high unemployment, the ripple effects can be pervasive. Research suggests that in some parts of the United States, factory closures contribute to a rise in deaths from opioid overdoses. Now imagine similar pressures on both white-collar and blue-collar workers nationwide.

We have to think now about how to reduce job losses so that everyone can share in the prosperity that AI creates. Waiting until people are already displaced or underemployed will be too late. AI is a structural challenge to the way our economy is organized, and it requires thinking and action now.

AI will empower people (and perhaps AIs) to do more harm.

Long before AI entered the mainstream, there was information online about how to create weapons like bombs, bioweapons, even computer viruses. AI will make it much easier to not only get this information but act on it. Even criminals with very limited skills will be able to target victims at every scale: individuals, companies, and governments.

Even criminals with very limited skills will be able to target victims at every scale.

AI-enabled fraud, disinformation, deepfakes, and surveillance are the harms that many people will feel most keenly in their everyday lives.

AI capabilities are starting to be used for cyberattacks. The smartest cybersecurity experts I know are scared about the next few years, because the attackers are getting powerful new capabilities faster than the defenders can fix all the weaknesses. After all, the same AI model that can find a flaw in software so a company can fix it can also help a criminal exploit it. The resources needed to make an attack are going down significantly and we haven’t been able to separate those abilities from benign usage.

Think about the infrastructure that will be vulnerable: hospitals, financial institutions, water systems, power grids, systems for managing government benefits. When these institutions are attacked, it’s the patients, customers, and benefits recipients who stand to lose.

The same goes for bioterrorism. Although AI will lead to lifesaving advances in drugs and vaccines, it will also make it easier to design a deadly new disease. Again, the positive capabilities are hard to separate from the dangerous ones. This is a global problem.

The risks I’ve just mentioned are all about how AI will empower bad actors who have relatively little power now. The same tools will also concentrate power in places where it already exists. Autonomous weapons, for example, will make governments even more capable of using deadly force without a human being part of the decision. Monitoring and manipulating public opinion will be easier and cheaper, and more effective too.

Eventually, the power to use AI to harm people will not be limited to people or institutions. AI systems themselves already occasionally act in ways their designers didn’t intend. The technology is improving faster than anyone expected and in surprising ways, and as the models become more powerful, they could begin to act against our interests and we could lose control. I’ll have more to say about this in the future.

AI models could begin to act against our interests and we could lose control.

AI could stunt our kids’ development and replace human relationships.

When I was growing up in Seattle, I didn’t have that many friends aside from a few other boys who were like me. It took hard work and a lot of help from my mom to develop my social skills so I could relate to different kinds of people. I still draw on those lessons today at the age of 70.

I doubt I would have put in the same work if I had had an AI companion back then. They talk to you in ways you’re already comfortable with. They don’t push you outside your comfort zone. They are always available and never get mad at you. This gives them the potential to become highly addictive and to rob us of the lessons we learn from connecting with other people.

The body of evidence on this subject is still small and a bit mixed, but there are signs that we should be very concerned. For example, in one study of more than 1,100 people who use AI companions, researchers at Stanford and Carnegie Mellon found that those with smaller social networks were the most likely to turn to a chatbot for companionship. And the heavier and more emotionally personal that use became, the worse they felt.

Young people could be affected for their entire lives. In his book The Anxious Generation, Jonathan Haidt makes an observation about the effect of social media that is even more true for AI: “Like young trees exposed to wind, children who are routinely exposed to small risks grow up to become adults who can handle much larger risks without panicking. Conversely, children who are raised in a protected greenhouse sometimes become incapacitated by anxiety before they reach maturity.”

An AI companion designed to never upset you is a big, protected greenhouse.

We are only beginning to understand the dangers that the internet—especially social media—can pose to young people’s development. We’re seeing compulsive use, disrupted sleep, cyberbullying, and exposure to harmful content. AI could magnify many of these risks by making them more persuasive and difficult to escape, and we should not wait another generation to start taking them seriously. Countries including Australia, the United Kingdom, and Norway are adopting protections for children online. China has gone the furthest. Its rules restrict AI companion apps broadly, bar designs that foster emotional dependence, and ban virtual relatives and romantic partners for minors.

The same tool that will allow people to learn more than ever could also lead to many people learning less.

I’m also worried about AI’s impact on education. Ironically, the same tool that will allow people to learn more than ever could also lead to many people learning less. One preliminary survey suggested that heavier AI use was associated with less critical thinking. The effect was stronger for younger people.

This would be the worst possible time for humans to lose their critical thinking skills. In an era of deepfakes and misinformation that can be tailored to you individually, the ability to tell what is true from what is not becomes an essential life skill.

It’s unclear where to draw the line on these psychosocial problems. In some cases, AI may help people understand how to do better in their human relationships. It may be the only contact with the outside world for isolated elderly people and people with limited mobility, and it will be better than nothing. Wherever we end up drawing the line, it should be our decision, made intentionally.

The good things we do with AI could be very, very good.

It’s often said that we overestimate how much will change in the short term and underestimate how much will change in the long term.

With AI, I see something different going on. Some people see only the upside of AI and do not focus enough on the negatives. Others make the opposite mistake, which is to focus exclusively on the dangers—which are real—at the cost of missing the potential benefits.

We need both: deep concern about the AI harms we need to minimize, and grounded optimism about the positives if we maximize them for everyone.

Maximizing the benefits is just as important as minimizing the harms. If people see how AI makes their lives easier, it will help build the public trust that is necessary for managing the harder parts of the transition. If the first thing AI does in most people’s lives is take away their job, those who are already skeptical about it will outright reject it. This will make it harder to ever deliver on the benefits and it is another reason why governments, industries including the medical industry, and AI companies should be working together now.

With its ability to synthesize knowledge from every scientific field, AI can accelerate innovation in the world’s toughest technical challenges: providing reliable clean energy for everyone, combating climate change, growing enough food, eradicating diseases, and more. Researchers working on cancer treatments or nuclear energy can use AI to search through massive amounts of scientific literature. It can help them identify patterns that a human might miss and decide which experiments offer the most promise. When intelligence is no longer the limiting factor that it is today, smaller companies will be able to compete with organizations that have far larger research budgets. R&D and innovation will be supercharged.

Healthcare is one area where AI can help solve real-world problems. Many small American hospitals lack on-site specialists who can quickly diagnose a patient during a life-threatening emergency. In those places, AI could make sure a heart attack is caught in time and a family avoids the crushing expense of a medical emergency. Viz.ai is one example. It analyzes scans to detect strokes and other emergencies and helps medical teams coordinate their patients’ care. It is being used in nearly 2,000 U.S. hospitals.

AI will also help primary-care doctors make better diagnoses and keep in touch with their patients when they’re not in the clinic. It will help patients understand test results and complicated schedules for taking their medicine.

Agriculture is where I see the fastest impact of AI in low-income countries.

I surprise a lot of people when I tell them that a second area—agriculture—is where I see the fastest impact of AI in low-income countries. In most low-income countries, farmers don’t get reliable weather forecasts or advice on what seeds to plant, how to protect their crops and livestock from disease, or how to improve their soil. With population growth in these countries and the challenges of climate change, these farmers need more help than ever. Using AI, low-income farmers will soon be able to get better advice about all these things than even the richest farmers get today and increase their output substantially.

Government services are a third area where AI can make people’s lives easier. In the United States, I’ve met families who, understandably, were overwhelmed by the process of applying for health insurance, student aid, or food assistance. Faced with a huge stack of complicated bureaucratic forms, many felt like giving up. AI can streamline things dramatically so they get the help they need faster and the government can operate more efficiently. Governments can make the citizen’s experience far better, starting with those who need its safety net services the most.

Despite my concerns about its impact on our mental health, I think AI can also help a lot there. Most communities have too few counselors, psychiatrists, and addiction specialists. With the right privacy safeguards in place, AI tools could help people recognize warning signs. Then, if needed, they can offer evidence-based coping strategies and team up with a human to provide more responsive treatment.

AI can be a boon for education as well, despite the concerns I mentioned earlier. It can free teachers up to spend more time working with students one on one or in small groups and give them a clearer view of where the whole class is struggling. For students, an AI tool that preserves what researchers call “productive struggle”—the cognitive work that builds understanding—can strengthen learning. When a student first encounters a new idea, the AI gives substantive explanations and offers both questions and answers. Later, when it’s checking their comprehension, it holds the answer back and helps them arrive at it on their own.

Taken together, the advances in all these areas could make everyday life easier, more affordable, and less constrained by a person’s income or connections.

AI could give individuals and small businesses access to capabilities that today require expensive professional help or large staffs, while making products and services better and cheaper. It could help people with disabilities live more independently and enable workers and entrepreneurs with good ideas to accomplish far more than they can today.

Most importantly, it could give people back some of the time and attention now consumed by paperwork, bureaucracy, searching for reliable information, and tasks they cannot afford to pay someone else to handle. These benefits may seem modest, but multiplied across millions of lives, they would be profound: more people getting good advice when they need it and having greater freedom to focus on the lives they want to build.

We have to be deliberate about ensuring that it benefits everyone and not just a wealthy few.

In all these areas, the operative word is “can”—AI can improve life for people at every income level. But it won’t do that automatically. As with any new technology, we have to be deliberate about ensuring that it benefits everyone and not just a wealthy few. This will require governments and philanthropy to play a strong role so that less wealthy citizens and low-income countries are full beneficiaries.

The Gates Foundation has 19 years left of the 20 years in which it will spend its remaining $200 billion. AI will help it achieve its ambitious goals by both accelerating the discovery of vaccines and medicines for HIV, TB, malaria, and malnutrition and helping the healthcare workforce and patients know how to use those tools. The foundation’s goals include cutting the number of children who die every year in half again, as was done from 2000 to 2024. All of our work, not just health but also agriculture and education, will take full advantage of AI.

I will write much more about these efforts next month in the foundation’s annual Goalkeepers report—including our focus on making sure that AI models are available in the languages spoken by people in all the countries where we support work, and not just the ones that are common in rich and middle-income countries. Many of the leading AI companies, including OpenAI, Anthropic, Google, and Microsoft, are partnering with the foundation on all of these initiatives, which is making a big difference.

The world needs a plan.

It is great that some AI companies are proposing solutions to challenges raised by their own technology, but we should not expect them to lead the charge. Some of the issues are outside their area of expertise, and in a democratic society it’s not their role to decide these things.

Instead, solutions should be developed through a public democratic process that includes elected officials, policymakers, educators, health workers, local officials, and community leaders. Millions of people will have their lives disrupted, and we’ll need a stronger, more flexible social safety net to help them manage the transition. Local communities are already raising concerns about the energy and water needed for data centers. Without solutions, some groups will push for stopping AI development and deployment altogether.

The solutions should be shaped by our answers to the profound questions raised by AI, including how we preserve our humanity in a time when machines can out-think us. As people who spend their lives thinking about what it means to be human, religious leaders can play a key role in this. I was fascinated by Pope Leo XIV’s encyclical on AI, “On Safeguarding the Human Person in the Time of Artificial Intelligence.” It lays a strong foundation for the work that needs to be done.

In the coming months, I will share more ideas for making sure that AI’s benefits outweigh the harm it causes. Here are three to start, beginning with what I think is the most important one.

Build a new system for managing the transition.

The highest priority is a monumental task: creating a domestic and international framework for dealing with AI.

None of our current institutions were designed to handle a technology that spreads so fast and touches so many parts of our lives. So we’ll need to make new ones.

It’s hard to overstate what an enormous undertaking this will be. After the attacks of 9/11, the U.S. government went through its biggest reorganization since World War II for the purpose of improving just one function, national security.

AI will require much, much more. It will affect national security as well as employment, education, taxation, energy, elections, air and water, public health, the financial system, law enforcement, transportation, public lands, and IT systems.

These sectors overlap in ways our existing bureaucracy is not designed to manage. A labor department may understand workforce disruption but not security risk. A business regulator may understand market concentration but not AI’s effects on children and teenagers. Left to themselves, institutions will see only one part of the system, while the consequences of AI will ripple across the entire system.

At the national level, countries will need bodies that can set priorities across government agencies. The goal will be to make sure that every risk is accounted for. Otherwise, an AI-enabled attack might succeed because no one thought it was their job to stop it.

But even a country that gets its own house in order will still be exposed to risks that cross borders. This is why an international organization will need to be built in parallel.

It will be unlike any other institution we have ever created, though it can follow the model of some existing systems. There’s an inspections regime for nuclear weapons, regulations for international aviation, and agreements that protect the ozone layer. A new global organization for AI will need elements of all three and more.

It is fair to wonder whether the world’s institutions are up to the task of designing and implementing this new architecture. Government moves slowly when it moves at all, and polarization within and between countries makes it harder than ever to get things done. Some cooperation between the U.S. and China will be required.

We do not have the luxury of moving slowly. The place to start is with a process for building the right institutions before the disruption forces governments into crisis mode. National leaders should convene economists, technologists, labor experts, business leaders, and workers themselves regularly to identify where existing institutions are failing and what new authorities may be needed. Countries will need to learn from each other.

And the countries that host the leading AI developers and control critical parts of the supply chain should begin meeting now to set up shared norms, before competitive pressure makes it harder for them to cooperate.

Building the framework I’m talking about will take years, which is why we need to start now.

Set aside some jobs for humans.

My dad died of Alzheimer’s in 2020. In the later stages of his illness, he was cared for day and night by paid caregivers who understood him even when he struggled to express himself. He couldn’t always tell them when he was hungry, but they always knew.

My family and I will always be grateful to that amazing group of professionals. Something in the care they gave my dad was irreplaceably human. No robot could or should have done it.

I think about that team when the question of which jobs will disappear and which will remain comes up. I believe that as AI and robots improve, we’ll set aside certain things for only people to do. I’ve started calling this domain Human Reserved, and it’s an example of the kinds of ideas we’ll need to consider.

I like the phrase Human Reserved because it makes me think of nature reserves—places where we could put buildings and roads, but we choose not to because the loss would be too great.

We might set something aside as Human Reserved for economic reasons. For example, we may do it because allowing machines to take over a certain role will displace a large number of people who can’t easily change jobs. You can’t tell a 55-year-old who has worked in construction their whole career that they need to go work at an elder care facility and expect them to find it fulfilling.

Sometimes the decision to make something Human Reserved will be driven by other factors. In health, for example, imagine a robot giving you the awful news that you have an incurable disease. There’s no technical reason why it couldn’t. Yet it shouldn’t.

The Human Reserved domain will evolve over time—for example, we should consider setting aside some jobs now and phasing in AI slowly over years or decades with a commitment to preserve some jobs. Some areas, like education and mental health care, will be a mix, with a human in charge who’s using the technology to extend what they can do.

The lines will also vary from place to place. Some countries might insist on having humans take care of the elderly. But a country like Japan, which has a shrinking workforce and not enough young people to care for the old, may welcome a caregiving robot.

The idea of Human Reserved raises a host of questions I don’t have answers to. Who gets to decide what we reserve for humans? What criteria should we use? How do you keep companies from cheating and using robots anyway? What happens to international trade when one country lets robots make something and another country doesn’t? These will need to be worked out in public as part of the transition plan.

Rebalance how we tax labor and capital.

As workers are pushed into different jobs, they will need retraining and other support from the social safety net. But they will be working less, which means they will be paying less in income taxes, and government revenues will drop just when the demand for those services is greatest. The funds will have to come from somewhere at a time when budgets are stretched.

I believe we should tax AI tokens and robots. Right now, if you’re an employer and you hire someone, you pay payroll taxes on their earnings. But if you buy a robot, you can usually write it off right away as a business expense. The tax system nudges you toward replacing people with machines.

A tax would slow the rush away from human labor a little and raise money for retraining and a stronger safety net. It would need to be targeted so it does not slow down the purely beneficial uses of AI, like making medicine and education cheaper.

Critics of this idea point out that it’s not optimally efficient in an economic sense, but they’re not considering the broader value of work for individuals and society. And with all the accelerated innovation we will have, we’ll be able to afford a little inefficiency as the price for keeping people employed.

I proposed a robot tax years ago and most of the reaction was that it was a strange idea. I’m still a big proponent of it. Although it is not the whole solution to the threat of AI, it is part of a wise response.

However we raise money for more assistance, it needs to reach the people who need it most, including workers who lose their jobs to AI and robots, people whose hours or wages decline, and communities where the losses are concentrated. We need to start doing that work now so that the systems are ready when the need becomes acute.

What I’m doing.

I will use my voice and time to get AI and equity higher on the public agenda. I will raise the issue with lawmakers every time I visit Washington, D.C., and when I meet with leaders around the world. It will be front and center in my conversations with the people who are developing AI models. I will advocate for the national and international framework I described earlier. The Gates Foundation will help drive beneficial usage, including in Africa. Breakthrough Energy, a company I founded, will use AI to help companies develop cheap clean energy and help solve the climate problem. I will also be writing about AI on a regular basis.

My message to leaders is:

You have a chance to act now, before unemployment rises sharply, communities are hurting, and public trust has eroded. You can make sure that your government handles the problem holistically, rather than divvying it up into multiple bureaucratic fiefdoms. You can make sure AI benefits everyone. And you can work with other governments to meet this national and global challenge.

Finally, I will try to widen the circle of people shaping this debate. It should include workers, college students who are about to enter the workforce, community leaders, religious leaders and faith-based organizations, parents, educators, and others whose voices often aren’t heard but who have insight into how the transition will affect people’s lives.

How do we ensure that the benefits of AI reach people who do not already have wealth, influence, and access?

How do we strengthen the social safety net and help workers and communities thrive even when they’re displaced?

How should public institutions adapt?

And how do we preserve our humanity through all of this?

This unprecedented technology demands an unprecedented global response.

This unprecedented technology demands an unprecedented global response. If we get it right, the payoff for humanity will be phenomenal and the world will be a more equitable place.

I rarely stop thinking about AI—not because I have all the answers, but because the questions it raises are too consequential to leave to a small group of technologists. Leaders across academia, business, government, and civil society all have a role to play in shaping what comes next.