Federal Reserve Governor Lisa D. Cook said artificial intelligence is likely to become the most significant technological shift of her lifetime, adding near-term inflation pressure while potentially driving productivity gains and a major reorganization of work.

Cook made the remarks on September 28, 2026, in an opening keynote at Oakland Tech Week, cohosted by the Kapor Center in Oakland, California. The full speech is available from the Federal Reserve at https://www.federalreserve.gov/newsevents/speech/cook20260928a.htm. As a member of the Federal Reserve Board of Governors, Cook holds a standing vote on the rate-setting Federal Open Market Committee, making her assessment of AI’s effects on prices and jobs directly relevant to the policy decisions ahead.

AI as a general-purpose technology

Cook said her interest in the economics of innovation and artificial intelligence began while studying at Berkeley, where she took a growth theory course taught by Paul Romer. Romer argued that investment in science and ideas could produce continuous, unbounded growth and later received the 2018 Nobel Memorial Prize in Economic Sciences for his contributions to the study of economic growth.

Cook said the current period, combined with the historic post-World War II investment in basic science, is seeing those ideas tested in the global economy. She described AI as a general-purpose technology comparable to, or potentially exceeding, breakthroughs such as the steam engine, electricity and the internet. Those earlier innovations spread across the economy, generated downstream innovation and improved over time; Cook said AI is following a similar pattern.

As a researcher of innovation economics and a monetary policymaker, Cook said she has been closely observing AI’s effects. The technology affects nearly every aspect of her work at the Federal Reserve, including monetary policy, financial stability, bank supervision, financial market infrastructure and the central bank’s operational preparedness.

Cook said AI presents substantial opportunities but also requires a cautious approach. She cited concerns involving the data used to train models, AI safety, privacy, bias, workers, fraud, cybersecurity and intellectual property rights. She said researchers, consumers, businesses and policymakers will determine whether AI’s potential benefits are realized and how those benefits are distributed.

As a supervisor of banking organizations, the Federal Reserve has encouraged and will continue to encourage responsible AI use in the financial sector, Cook said, provided that its use is consistent with safe and sound practices and applicable law. She added that any specific government role in determining AI’s trajectory is a matter for elected officials.

Cook said she would continue monitoring the speed, direction and magnitude of AI development as she assesses the economic outlook, the appropriate path of monetary policy and financial stability. Her remarks addressed AI’s implications for the Federal Reserve’s dual mandate of price stability and maximum employment, goals Congress has assigned to the central bank.

AI and inflation

Cook said inflation has remained above the Federal Reserve’s 2 percent target over the past five years for a broad range of reasons. Over the past year, one contributing factor has been AI-driven investment, which has pushed up prices for AI-related goods such as chips, computers and software.

Some of those increases, she said, reflect a shift in demand toward AI-related sectors rather than an increase in demand across the entire economy. When demand surges in a single sector, goods and services in that sector can move onto a steep portion of their supply curve. If demand were distributed evenly across the economy, the overall price index would not rise as much. Cook said supply-chain adjustments and efficiency gains should eventually ease this sector-specific pressure without monetary-policy intervention.

Using monetary policy to address inflation concentrated in a narrow sector could be a mistake, Cook said, because the Federal Reserve’s tools are too broad to target individual sectors and addressing relative price shifts is not the central bank’s role.

However, Cook also sees economy-wide pressure from AI-related demand. Data-center investment requires construction labor and energy, inputs that are used broadly across the economy. Increased AI investment could therefore place price pressure on other sectors. She noted that companies have spent only a small fraction of the $2 trillion in announced data-center investment plans, leaving more investment in the pipeline.

Cook also said a large portion of the increase in equity prices over the past few years can be attributed to enthusiasm about AI, with the resulting wealth apparently feeding into household spending. She pointed to signs that inflationary pressure may be broadening: electricity and water costs are each up around 5 percent over the past year, potentially partly because of AI, while core goods prices—which had been declining before the pandemic—are running at an annual pace above 3 percent so far this year.

That creates a risk that new, broader price pressures could replace inflation in the narrower AI sector even as sector-specific inflation moderates, Cook said.

A well-timed productivity boom could counter that pressure if it increases the economy’s supply capacity more than it increases demand. Lower input costs can allow businesses to produce more goods and services at the same cost. If demand does not rise enough to absorb the additional supply, prices would be expected to fall. Cook described this dynamic using the aggregate supply and aggregate demand framework.

Cook currently expects productivity gains to provide modest disinflation within the next few years, but she does not expect those effects to arrive soon enough to offset broader inflationary pressure later in 2026. She said uncertainty surrounds estimates of how and when the productivity mechanism will operate and that the issue requires additional research and discussion.

The extent of inflation relief will depend on whether a productivity boom increases the economy’s supply potential more than it increases demand. If productivity gains are distributed evenly, Cook expects the relief to be limited but real. Higher productivity raises potential supply, but it can also increase demand through expectations of higher future wages, better investment returns and related wealth gains. Some of that additional demand may already be reflected in the economy through data-center investment and wealth effects.

Going forward, new supply generated by higher productivity could exceed new demand, bringing supply and demand into better alignment and reducing current upward pressure on prices. On balance, Cook said, a broad-based productivity increase, if and when it occurs, could produce a modest easing in price pressure.

She also outlined scenarios that could produce either more or less disinflation than her baseline. Greater inflation could occur if demand fell significantly below the economy’s capacity. That could happen if productivity gains were concentrated among higher-income consumers, who tend to spend less of each additional dollar of income or wealth; if weak worker bargaining power prevented productivity gains from being passed through to wages; or if a difficult period of job reallocation increased unemployment and reduced confidence, leading households to increase precautionary savings.

Less inflation relief could result if productivity gains were not passed through to prices, for example because market concentration reduced competition and enabled higher markups.

The timing of any disinflationary benefit will ultimately depend on how quickly and broadly businesses adopt AI tools, how changes in business practices translate into higher productivity and how rapidly productivity gains pass through to the labor market. Cook said she remains highly uncertain about the breadth and timing of those channels and will adjust her assessment based on the data.

AI and the labor market

Cook said AI’s potential effects on the labor market have long attracted public attention and warrant close study by policymakers. Historical technologies that changed how people work, including the mechanical loom and the personal computer, generated widespread anxiety when introduced. The benefits and costs of technological change have not always arrived at the same time, she said, and the same could occur with AI.

AI could bring the most significant reorganization of work in generations, Cook said. However, the technology and its adoption remain in their early stages, and there is currently limited evidence of significant changes to the overall structure of the labor market. The unemployment rate and layoffs remain low and have been relatively flat over the past two years even as AI adoption has increased.

Cook said she has been encouraged by the labor market’s resilience during the early stages of AI adoption. At the same time, some workers have already been affected, and she acknowledged the hardship that unemployment creates for workers and their families.

There is evidence in certain sectors, including software-industry coding and simultaneous translation, that AI may be reducing labor demand. Many recent college graduates are also having more difficulty finding their first jobs, possibly partly because AI can perform entry-level tasks. Cook said these shifts could broaden as AI adoption deepens and the technology improves.

AI is also likely to generate new innovation, tasks and occupations that cannot currently be anticipated. Cook said she expects AI to fundamentally change business practices and transform the labor market in a corresponding way. She expressed hope that AI adoption will proceed at a pace and in a manner that allows job creation to match or exceed job destruction.

Cook remains attentive to a scenario in which AI causes at least a temporary increase in unemployment. In such a case, higher unemployment might not reflect increased economic slack or insufficient demand, but rather a supply-side mismatch between workers’ skills and available jobs. The Federal Reserve would have limited tools in that situation. It could lower the federal funds rate to try to reduce unemployment, but doing so could risk adding to inflation.

As a policymaker, Cook said she is monitoring the speed and acceleration of AI adoption across sectors, evidence of changes in the mix of jobs within firms and whether new job creation is offsetting displacement. She said she evaluates a broad set of data, the evolving outlook and the balance of risks when making policy decisions. Fiscal policy could also help manage AI’s effects on the labor force, she said, but that falls outside the Federal Reserve’s remit.

AI adoption by smaller firms

Cook said she is studying how AI affects firms of different sizes. Although rapidly growing, high-profile companies lead in AI technology and other large technology companies have made substantial investments, she challenged the view that only large firms with significant resources can deploy advanced AI.

She said smaller companies are showing resourcefulness and adaptability in using AI, even though they lag larger firms in overall adoption. Early-stage investors and founders across the country have told her that AI tools have made it easier and cheaper to start businesses. Cook said that could be contributing to the recent historic boom in new business formation and could allow AI to benefit smaller companies and their workers.

The Federal Reserve Small Business Credit Survey—a collaboration among the 12 regional Federal Reserve banks—released earlier in 2026 found that nearly half of small employer firms are using AI and that 71 percent report increased productivity as a result. Cook said those findings indicate that many small businesses are actively seeking innovative tools to manage and expand their operations. Greater use of those tools could help small firms compete more effectively, grow and create jobs.

Evidence also shows that AI adoption is occurring faster than personal-computer or internet adoption at comparable stages, Cook said. That is significant for small businesses because some of the country’s most important breakthroughs have come from start-ups using frontier technologies in new ways.

Small businesses with fewer than 500 employees represent more than 99 percent of all U.S. businesses and have accounted for 61 percent of net new job creation since 1995. Cook said that when those firms become more productive, the broader economy is strengthened.

When powerful tools are broadly available rather than limited to large corporations, more inventors, innovators and entrepreneurs can use them, Cook said. She connected that possibility to Paul Romer’s view that long-term economic growth is generated within the economy by people creating new ideas and knowledge. AI could help people generate more ideas more quickly, suggesting that the economy may be experiencing what she called a “Romer moment.”

If AI adoption becomes as broad-based as Cook believes possible and competition among model producers remains robust, productivity gains will not necessarily be concentrated among a few companies. In that case, AI could help small companies scale more quickly and create more jobs, although Cook emphasized that this is not guaranteed.

Cook said AI will transform the economy, potentially on a larger scale than previous general-purpose technologies. The effects of those earlier breakthroughs, however, often involved long and variable lags, a concept familiar to monetary policymakers. The full benefits of AI will depend on decisions by firms, workers and policymakers, including investment not only in chips and data centers but also in worker training, organizational changes and the creation of new processes and ideas.

Cook described the implications over three time horizons. In the short term, AI appears to be adding inflationary pressure and delaying inflation’s return to the Federal Reserve’s 2 percent target. In the medium term, productivity growth may modestly reduce those pressures, but the labor market could face a painful transition. In the long term, Cook said she is optimistic that AI-driven productivity growth can raise living standards for all Americans.

With vigilance and careful management of the opportunities and challenges, Cook said, the Federal Open Market Committee will be able to achieve its dual mandate of maximum employment and stable prices.

Economic outlook and monetary policy

Cook concluded by discussing her broader economic outlook. She said she voted with the rest of the FOMC to raise interest rates by 25 basis points—a quarter of a percentage point—at its recent September meeting. The increase was intended to address inflation that has remained too high for too long.

Total inflation rose an estimated 3.8 percent in the 12 months through August, nearly double the Federal Reserve’s target. Core inflation, which excludes energy and food prices, rose an estimated 3.4 percent. Cook said she expects continued inflation pressure in coming months from the AI buildout as well as from the pass-through of higher oil prices and supply-chain disruptions related to the conflict in the Middle East.

She said the labor market appears well positioned to handle higher interest rates. The unemployment rate had been trending downward during the year, with the latest reading at 4.1 percent in August. Other indicators showed a labor market that was roughly in balance and gradually improving: payrolls had increased, job openings had risen and initial unemployment claims had trended lower.

The broader economic-growth data also showed a remarkably resilient economy over the past year, Cook said. She will assess what policy rate may be needed to continue guiding inflation toward the Federal Reserve’s target. The number and size of any future adjustments will depend on how the economy responds to policy actions already taken and on inflation and labor-market data over the coming months.

References

1. Cook said the views expressed in the speech were her own and were not necessarily those of her colleagues on the Federal Reserve Board or the Federal Open Market Committee.

2. Lisa D. Cook, “Generative AI, Productivity, the Labor Market, and Choice Behavior,” speech delivered at the National Bureau of Economic Research Economics of Artificial Intelligence Conference in Toronto, Canada, September 22, 2023: https://www.federalreserve.gov/newvents/speech/cook20230922a.htm.

3. Lisa D. Cook, “The Opportunities and Risks AI Presents for the Economy and Financial System,” speech delivered at the Stanford Institute for Economic Policy Research at Stanford University in Stanford, California, May 27, 2026: https://www.federalreserve.gov/newsevents/speech/cook20260527a.htm.

4. Lisa D. Cook, “Artificial Intelligence, Big Data, and the Path Ahead for Productivity,” speech delivered at “Technology-Enabled Disruption: Implications of AI, Big Data, and Remote Work,” a conference organized by the Federal Reserve Banks of Atlanta, Boston and Richmond in Atlanta, Georgia, October 1, 2024: https://www.federalreserve.gov/newsevents/speech/cook20241001a.htm.

5. Eirik Eylands Brandsaas, Daniel Garcia, Robert Kurtzman, Joseph Nichols and Adelia Zytek, “Estimating Aggregate Data Center Investment with Project-Level Data,” Finance and Economics Discussion Series 2025-109, Board of Governors of the Federal Reserve System, December 2025. Updated data and publicly available results are available from Eirik Brandsaas, “Estimating Aggregate Data Center Investment with Project-Level Data,” DataCenterPublic, GitHub repository: https://github.com/eirikbrandsaas/DataCenterPublic.

6. Lisa D. Cook, “Opening Remarks for the ‘AI and Productivity across the Economy’ Panel,” speech delivered at “The Great Realignment: Navigating AI, Demographic, and Geoeconomic Shifts,” the 42nd Annual NABE Economic Policy Conference in Washington, D.C., February 24, 2026: https://www.federalreserve.gov/newsevents/speech/cook20260224a.htm.

7. Lisa D. Cook, “Welcome Remarks,” speech delivered via prerecorded video at the State of Small Business Symposium hosted by the Federal Reserve Bank of Cleveland, June 24, 2026: https://www.federalreserve.gov/newsevents/speech/cook20260624a.htm.

8. Alexander Bick, Adam Blandin and David J. Deming, “The Rapid Adoption of Generative AI,” NBER Working Paper Series 32966, National Bureau of Economic Research, September 2024; revised February 2025:

9. Paul M. Romer, “Endogenous Technological Change,” Journal of Political Economy, volume 98, October 1990, pages S71–S102.