AI and the Economy
Thank you, President Nagel and the Bundesbank, for the opportunity to meet with you today.
感谢内格尔行长和德国央行给我今天与你们会面的机会。
1 I am excited to be here.
1 我很高兴来到这里。
I cherish opportunities to talk with groups of young people such as this.
我很珍惜与这样一群年轻人交谈的机会。
I spent most of my career as a professor and university administrator.
我职业生涯的大部分时间都是作为教授和大学管理人员度过的。
I am glad to return to those roots a bit today.
我很高兴今天能稍微回到这些根源。
I will make sure I leave plenty of time for questions.
我会确保留出足够的时间来提问。
But I would like to start by speaking with you about a topic that's been on my mind, and I bet it has been on many of yours: artificial intelligence (AI).
但我想首先与大家讨论一个我一直在思考的话题，我敢打赌这也是你们很多人的想法：人工智能 (AI)。
Often, when I gather with students, consumers, or businesspeople, the conversation quickly turns to AI.
通常，当我与学生、消费者或商人聚会时，谈话很快就会转向人工智能。
And it is easy to understand why—you do not even need to ask a chatbot!
原因很容易理解——您甚至不需要询问聊天机器人！
AI has the potential to transform how we work and how we live, and it could have a profound effect on the economy.
人工智能有潜力改变我们的工作和生活方式，并且可能对经济产生深远的影响。
Today, I will talk about AI's rapid growth and how I approach thinking about AI by examining it through the lens of the dual mandate given to the Federal Reserve by the U.S. Congress.
今天，我将讨论人工智能的快速发展，以及我如何通过美国国会赋予美联储的双重使命来审视人工智能。
That mandate instructs America's central bank to pursue maximum employment and price stability.
该指令指示美国央行追求就业最大化和价格稳定。
I will describe the effects that AI may have on both sides of that mandate.
我将描述人工智能可能对该任务双方产生的影响。
Finally, I will briefly update you on my economic outlook before turning to your questions.
最后，在回答你们的问题之前，我将简要介绍一下我的经济前景。
AI's Rapid Growth In a few short years, AI morphed from the realm of science fiction to a tool that high schoolers use to help them do their science homework!
人工智能的快速发展 在短短几年内，人工智能从科幻小说的领域变成了高中生用来帮助他们做科学作业的工具！
Seriously, it is amazing to see school children use AI to generate quizzes to prepare for exams, home chefs use it to create recipes for a wholesome dinner, and jobseekers use it to consider career tradeoffs, among many other purposes.
说真的，看到学童使用人工智能生成测验来准备考试，家庭厨师使用它来制作健康晚餐的食谱，求职者使用它来考虑职业权衡以及许多其他目的，真是令人惊奇。
Consumer use of AI has grown rapidly, and it is integrated into a multitude of existing products, including search engines, phones, and vehicles.
消费者对人工智能的使用迅速增长，并且已集成到多种现有产品中，包括搜索引擎、手机和车辆。
As of last month, ChatGPT, which is just one of a growing number of generative AI models, is used by more than 800 million people every week, an increase from 500 million a week at the end of March.
截至上个月，ChatGPT 只是数量不断增加的生成式 AI 模型之一，每周有超过 8 亿人使用，比 3 月底的每周 5 亿人有所增加。
2 Enterprise use of AI has been growing too.
2 企业对人工智能的使用也在不断增长。
A recently released study showed that the proportion of Americans using AI at their jobs has also been growing at a fast clip.
最近发布的一项研究表明，美国人在工作中使用人工智能的比例也在快速增长。
Researchers found that workers' adoption of generative AI rose from 30.1 percent as of December 2024 to 45.9 percent as of June and July of this year.
研究人员发现，员工对生成式人工智能的采用率从 2024 年 12 月的 30.1% 上升到今年 6 月和 7 月的 45.9%。
3 And among survey respondents who had adopted AI tools, about a third said they use them daily.
3 在采用人工智能工具的调查受访者中，约三分之一的人表示他们每天都使用这些工具。
That same study found that AI is most used at jobs by younger, more highly educated, and higher-earning workers and that productivity for workers increased substantially when using AI tools.
同一项研究发现，人工智能在工作中最常被年轻、受过高等教育、收入较高的工人使用，并且使用人工智能工具时，工人的生产力大幅提高。
There is strong evidence that AI can increase labor productivity in various industries and occupations.
有强有力的证据表明人工智能可以提高各个行业和职业的劳动生产率。
For example, a paper from Stanford economist, Erik Brynjolfsson, and coauthors showed the effects of AI implementation among customer support agents.
例如，斯坦福大学经济学家 Erik Brynjolfsson 及其合著者的一篇论文展示了人工智能在客户支持代理中实施的效果。
4 They found that AI tools increased productivity—as measured by issues resolved per hour—14 percent, on average, and the improvement in productivity for novice and low-skilled employees was much larger.
4 他们发现，人工智能工具平均提高了 14% 的生产力（按每小时解决的问题来衡量），而新手和低技能员工的生产力提高幅度要大得多。
AI is also advancing productivity in highly technical scientific research.
人工智能还提高了高技术科学研究的生产力。
Take protein folding, for example.
以蛋白质折叠为例。
Proteins are the building blocks of the human body and life itself.
蛋白质是人体和生命本身的组成部分。
They assume an almost infinite variety of shapes.
它们的形状几乎有无限多种。
Predicting those shapes—which determine their function—is vital for understanding how cells and diseases work and for finding drugs to treat them.
预测这些形状（决定它们的功能）对于理解细胞和疾病的工作原理以及寻找治疗它们的药物至关重要。
Just five years ago, scientists had a good understanding of the structures of just 17 percent of the roughly 20,000 proteins in the human body.
就在五年前，科学家们对人体大约 20,000 种蛋白质中的 17% 的结构有了很好的了解。
5 For each protein, that knowledge was earned through months or years of hard work and a potential cost of tens of thousands of dollars or more.
5 对于每种蛋白质，这些知识都是通过数月或数年的辛勤工作以及数万美元或更多的潜在成本获得的。
Thanks to the development of the AlphaFold model, an AI system that predicts a protein's structure, scientists now know the structures of all the proteins in the human body—and some 200 million more.
由于 AlphaFold 模型（一种可预测蛋白质结构的人工智能系统）的开发，科学家们现在了解了人体中所有蛋白质的结构，以及大约 2 亿种蛋白质的结构。
6
6
More broadly, like many others, I see the promise for AI technology to be transformative, perhaps in ways as dramatic as past technological achievements, such as the printing press, steam engine, and the Internet, which are known as general purpose technologies for their wide-ranging effects.
更广泛地说，像许多其他人一样，我看到人工智能技术有望带来变革，也许会像过去的技术成就一样引人注目，例如印刷机、蒸汽机和互联网，这些技术因其广泛的影响而被称为通用技术。
Labor Market and Prices When I think about this technology, I consider how it will affect both sides of the Fed's dual mandate of promoting maximum employment and stable prices.
劳动力市场和价格 当我思考这项技术时，我会考虑它将如何影响美联储促进就业最大化和稳定价格的双重使命。
As such, it is something that economists and policymakers must monitor.
因此，经济学家和政策制定者必须对此进行监控。
AI can allow a worker to complete tasks in moments that previously took many minutes, if not hours.
人工智能可以让工人在短时间内完成以前需要几分钟甚至几小时的任务。
That has caused many to question whether AI will lead to notable job loss.
这让许多人质疑人工智能是否会导致大量失业。
This potential disruption of labor is a real risk.
这种潜在的劳动力中断是一个真正的风险。
By automating certain tasks, it could lead to a reduction in some types of jobs.
通过自动化某些任务，可能会导致某些类型的工作减少。
But increased productivity leads to economic growth, which may create new employment opportunities.
但生产率的提高会带来经济增长，从而可能创造新的就业机会。
AI is also expected to create new job categories and transform existing ones.
人工智能还有望创造新的工作类别并改变现有的工作类别。
There has been robust competition among high-tech firms for workers who possess the skills to develop this technology.
高科技公司之间对拥有开发这项技术技能的工人的竞争非常激烈。
7
7
Many have legitimate concerns that AI will cause job loss.
许多人担心人工智能会导致失业，这是合理的。
At least for certain firms and workers in certain occupations, this is likely to be true.
至少对于某些公司和某些职业的工人来说，这可能是真的。
There are many examples where transformative technologies, such as the personal computer, did cause some workers to lose their jobs, but we also saw new industries and occupations emerge because of the innovation.
有很多例子表明，个人电脑等变革性技术确实导致一些工人失业，但我们也看到新的行业和职业因创新而出现。
Indeed, some large employers have recently indicated they are lowering overall hiring plans in light of advances in AI and other forms of automation.
事实上，一些大型雇主最近表示，鉴于人工智能和其他形式的自动化的进步，他们正在降低总体招聘计划。
8 Some research has also suggested that AI is having a more detrimental effect on the job prospects of younger, less experienced workers, including recent graduates, relative to those further along in their careers.
8 一些研究还表明，相对于那些职业生涯较晚的人来说，人工智能对年轻、经验不足的工人（包括应届毕业生）的就业前景产生更不利的影响。
9 The net effect on employment is highly uncertain and may show variation across industries and occupations over time.
9 对就业的净影响高度不确定，并且随着时间的推移，不同行业和职业之间可能会出现差异。
For the overall economy, one way to think about AI-related job losses—at least until it creates new jobs—is that it can create economic slack.
对于整体经济而言，思考与人工智能相关的失业的一种方法是，至少在它创造新的就业机会之前，它会造成经济疲软。
In a world where firms only deployed AI to replace existing workers, even at a lower level of overall employment, the productive capacity of the economy has increased (or stayed the same).
在一个企业仅部署人工智能来取代现有工人的世界中，即使总体就业水平较低，经济的生产能力也会增加（或保持不变）。
Of course, if people seamlessly get redeployed to more productive tasks, the economy produces more without any additional slack—more on this soon.
当然，如果人们无缝地重新部署到更有生产力的任务，经济就会在没有任何额外闲置的情况下产生更多的产出——很快就会有更多的内容。
Whether AI substitutes for labor or complements it is an important question that requires continued study.
人工智能到底是替代劳动力还是补充劳动力，是一个需要持续研究的重要问题。
On the other side of our dual mandate, price stability, AI could help the economy achieve higher growth through increased productivity while reducing inflationary pressures.
另一方面，我们的双重任务——价格稳定——人工智能可以通过提高生产率来帮助经济实现更高的增长，同时降低通胀压力。
Specifically, increased productivity could lower production costs and put downward pressure on prices.
具体来说，生产率的提高可以降低生产成本并对价格造成下行压力。
Not only could AI allow a specific worker to be more efficient in their day-to-day tasks, but it also has the potential to allow for better thinking around complex problems—for example, in logistics and transportation.
人工智能不仅可以让特定的工作人员在日常任务中更加高效，而且还有可能更好地思考复杂的问题，例如物流和运输方面的问题。
More efficient allocation of resources as well as potential improvements in supply chains could reduce associated costs, leading to lower prices for various goods and services.
更有效的资源配置以及供应链的潜在改进可以降低相关成本，从而降低各种商品和服务的价格。
Conversely, AI could put upward pressure on certain price categories as many firms push to scale up the technology.
相反，随着许多公司推动扩大技术规模，人工智能可能会给某些价格类别带来上行压力。
I mentioned the rising wages for certain workers with skills that complement AI.
我提到了某些具有与人工智能互补的技能的工人的工资上涨。
AI technology also requires data centers, which compete with other production processes for land, energy, and other inputs.
人工智能技术还需要数据中心，数据中心与其他生产流程争夺土地、能源和其他投入。
So, I think that AI's effect on inflation is not solely downward pressure.
所以，我认为人工智能对通胀的影响不仅仅是下行压力。
Implications for monetary policy So, what do these still-developing labor and price effects mean for monetary policy?
对货币政策的影响 那么，这些仍在发展的劳动力和价格效应对货币政策意味着什么？
The short answer is that it is likely still too soon to tell.
简而言之，现在下结论可能还为时过早。
As always, policymakers confront the challenge of sorting out changes in the economy that are due to cyclical factors from those resulting from structural change, which AI may well represent.
与往常一样，政策制定者面临着将周期性因素引起的经济变化与结构性变化（人工智能可能很好地代表）引起的经济变化区分开来的挑战。
Some of the recent changes in hiring patterns, productivity growth, and inflation are likely to represent AI-driven change, but it is difficult to know the degree.
最近招聘模式、生产率增长和通货膨胀方面的一些变化可能代表人工智能驱动的变化，但很难知道其程度。
Productivity gains from AI may affect the relationship between employment and inflation and hence the conduct of monetary policy.
人工智能带来的生产力提升可能会影响就业与通胀之间的关系，从而影响货币政策的实施。
For example, a productivity-induced boost to the growth rate of potential output could imply that monetary policy will not need to react strongly to what would have previously been perceived as tightness in the labor market.
例如，生产率引起的潜在产出增长率的提高可能意味着货币政策不需要对此前被视为劳动力市场紧张的情况做出强烈反应。
The uncertain implications of AI for employment and inflation could take some time to filter broadly through economy.
人工智能对就业和通胀的不确定影响可能需要一段时间才能广泛渗透到经济中。
Also, AI's effects will likely vary across different industries and regions of both the U.S. and the world.
此外，人工智能的影响可能会因美国和世界不同行业和地区的不同而有所不同。
And monetary policy decisions need to be made with a broad view of what is happening in the economy, not just in one sector or technology.
制定货币政策决策时需要广泛考虑经济中正在发生的事情，而不仅仅是某一部门或技术。
There is still much to learn.
还有很多东西需要学习。
I counsel exercising humility about the challenges in predicting AI's effects on employment and inflation.
我建议对预测人工智能对就业和通胀影响的挑战保持谦逊。
Policymakers, businesses, and individuals should remain flexible and prepared to adapt as the effects of AI continue to unfold.
随着人工智能影响的不断显现，政策制定者、企业和个人应保持灵活性并做好适应的准备。
Economic Outlook Now having discussed the policy implications of AI, I would like to briefly update you about my current outlook for the U.S. economy.
经济展望 现在讨论了人工智能的政策影响，我想向您简要介绍一下我目前对美国经济的展望。
While federal statistical agencies are largely not producing data during the government shutdown, data available from other sources suggest that the overall economic picture in the U.S. has not changed much over the past few months.
虽然联邦统计机构在政府关门期间基本上没有提供数据，但其他来源的数据表明，过去几个月美国的整体经济状况没有太大变化。
The economy has been growing at a moderate pace while labor market conditions appear to be consistent with a gradually cooling.
经济一直在温和增长，而劳动力市场状况似乎与逐渐降温相一致。
Inflation is running at a rate similar to that of a year ago.
通货膨胀率与一年前相似。
The lack of progress on headline inflation appears to be due to tariff effects, with signs that underlying inflation may be continuing to make progress to 2 percent.
总体通胀率缺乏进展似乎是由于关税效应，有迹象表明潜在通胀率可能继续升至 2%。
Market-based long-term inflation expectations continue to be well-anchored.
基于市场的长期通胀预期继续得到充分锚定。
I remain firmly committed to returning inflation to the Fed's 2 percent target.
我仍然坚定地致力于将通胀率恢复到美联储 2% 的目标。
I supported last week's decision to reduce our policy rate by a quarter percentage point.
我支持上周将政策利率降低四分之一个百分点的决定。
That step was appropriate because I see the balance of risks as having shifted in recent months as downside risks to employment have increased.
这一步骤是适当的，因为我认为近几个月来，随着就业下行风险的增加，风险平衡发生了变化。
The current policy stance is still somewhat restrictive, but we have moved it closer to its neutral level that neither restricts nor stimulates the economy.
目前的政策立场还是有一定的限制性，但已经接近不限制也不刺激经济的中性水平。
Given this, it makes sense to proceed slowly as we approach the neutral rate.
鉴于此，当我们接近中性利率时，缓慢推进是有意义的。
With respect to the path of the policy rate going forward, I will continue to determine policy based on the incoming data, the evolving outlook, and the balance of risks.
关于未来政策利率的路径，我将继续根据传入数据、不断变化的前景和风险平衡来确定政策。
I always take a meeting-by-meeting approach.
我总是采取逐次会议的方式。
This approach is especially prudent because it is unclear how much official data we will have before our December meeting.
这种方法尤其谨慎，因为尚不清楚在 12 月会议之前我们将获得多少官方数据。
Before I turn to your questions, let me conclude by saying that this is an exciting time.
在回答大家的问题之前，请允许我最后说，这是一个激动人心的时刻。
It feels like we are at the dawn of a new age—an age your generation will shape and see come to its full maturity.
感觉我们正处于一个新时代的黎明——你们这一代人将塑造并见证这个时代的完全成熟。
By the end of your careers, many of you will take on tasks and accomplish feats you cannot even imagine at this moment, and AI technology may well be part of that progress.
在你们职业生涯结束时，你们中的许多人将承担起目前无法想象的任务并完成壮举，而人工智能技术很可能是这一进步的一部分。
Thank you again for having me.
再次感谢您邀请我。
It is an honor to be here.
很荣幸来到这里。
I am looking forward to our discussion.
我期待着我们的讨论。
1.
1.
The views expressed here are my own and are not necessarily those of my colleagues on the Federal Reserve Board or the Federal Open Market Committee.
这里表达的观点是我自己的观点，不一定是我在美联储委员会或联邦公开市场委员会的同事的观点。
Return to text
返回正文
2.
2.
Rebecca Bellan (2025), "Sam Altman Says ChatGPT Has Hit 800m Weekly Active Users," TechCrunch , October 6, https://techcrunch.com/2025/10/06/sam-altman-says-chatgpt-has-hit-800m-weekly-active-users/. Return to text
Rebecca Bellan (2025)，“Sam Altman 称 ChatGPT 已达到 8 亿每周活跃用户”，TechCrunch，10 月 6 日，https://techcrunch.com/2025/10/06/sam-altman-says-chatgpt-has-hit-800m-weekly-active-users/。返回正文
3.
3.
Jonathan Hartley, Filip Jolevski, Vitor Melo, and Brendan Moore (2024), "The Labor Market Effects of Generative Artificial Intelligence," unpublished paper, available at SSRN: http://dx.doi.org/10.2139/ssrn.5136877. Return to text
Jonathan Hartley、Filip Jolevski、Vitor Melo 和 Brendan Moore (2024)，“生成人工智能的劳动力市场影响”，未发表的论文，可在 SSRN 获取：http://dx.doi.org/10.2139/ssrn.5136877。返回正文
4.
4.
Erik Brynjolfsson, Danielle Li, and Lindsey R.
埃里克·布林约尔松 (Erik Brynjolfsson)、丹妮尔·李 (Danielle Li) 和林赛·R.
Raymond (2023), " Generative AI at Work ," NBER Working Paper Series 31161 (Cambridge, Mass.: National Bureau of Economic Research, April; revised November 2023).
Raymond (2023)，“工作中的生成式人工智能”，NBER 工作论文系列 31161（马萨诸塞州剑桥：国家经济研究局，4 月；2023 年 11 月修订）。
Return to text
返回正文
5.
5.
Eduard Porta-Pardo, Victoria Ruiz-Serra, Samuel Valentini, and Alfonso Valencia (2022), "The Structural Coverage of the Human Proteome Before and After AlphaFold," PLoS Comput Biol ., vol. 18 (1), January 24.
Eduard Porta-Pardo、Victoria Ruiz-Serra、Samuel Valentini 和 Alfonso Valencia (2022)，“AlphaFold 前后人类蛋白质组的结构覆盖”，公共科学图书馆计算生物学，卷。 18 (1)，1 月 24 日。
Return to text
返回正文
6.
6.
EMBL-EBI (2023), "AlphaFold Protein Structure Database," webpage, https://alphafold.ebi.ac.uk/. Return to text
EMBL-EBI (2023)，“AlphaFold 蛋白质结构数据库”，网页，https://alphafold.ebi.ac.uk/。返回正文
7.
7.
Meghan Bobrowsky, Berber Jin, and Ben Cohen (2025), "Zuckerberg Leads AI Recruitment Blitz Armed With $100 Million Pay Packages," Wall Street Journal , June 22.
Meghan Bobrowsky、Berber Jin 和 Ben Cohen (2025)，“扎克伯格以 1 亿美元的薪酬包领导人工智能招聘闪电战”，《华尔街日报》，6 月 22 日。
Return to text
返回正文
8.
8.
For example, see Tom Huddleston Jr. (2025), "Walmart CEO: 'AI Is Literally Going to Change Every Job'—How The Best Employees Can Still Stand Out," CNBC, September 29; and Karen Weise (2025), "Amazon Plans to Replace More Than Half a Million Jobs With Robots," New York Times , October 21.
例如，请参阅 Tom Huddleston Jr. (2025)，“沃尔玛首席执行官：‘人工智能确实会改变每一份工作’——最优秀的员工如何仍能脱颖而出”，CNBC，9 月 29 日；和 Karen Weise (2025)，“亚马逊计划用机器人取代超过 50 万个工作岗位”，《纽约时报》，10 月 21 日。
Return to text
返回正文
9.
9.
J.P.
J.P.
Morgan (2025), "AI's Impact on Job Growth," webpage, https://www.jpmorgan.com/insights/global-research/artificial-intelligence/ai-impact-job-growth. Return to text
Morgan (2025)，“人工智能对就业增长的影响”，网页，https://www.jpmorgan.com/insights/global-research/artificial-intelligence/ai-impact-job-growth。返回正文