What Work Does Generative AI Do?
Essays SEP 8, 2026
By: Alexander Bick, Adam Blandin, David Deming, Tyler Schumacher
MAY 18, 2026
By: Jacob Robbins
Since regaining the Oval Office in early 2025, President Donald Trump has repeatedly called for lower interest rates1 and severely criticized former Federal Reserve Board Chair Jerome Powell’s cautious approach to interest rate policy. President Trump’s views culminated in the nomination of former Fed governor Kevin Warsh to replace Powell, explicitly linking his selection to Warsh’s public preference for lower interest rates.2
Both the president and Warsh have explicitly stated economic theories to support their preferred interest rate policies. President Trump has repeatedly asserted that lower interest rates will lead to higher economic output and thus economic “growth without inflation,”3 and that there is no meaningful inflation trade-off from easier monetary policy. In support of his claim, he offers two arguments. First, he argues that demand-driven growth, by expanding output, can itself fight inflation.4 And second, he says that his supply-side economic policies, such as deregulating fossil fuels and implementing high tariffs, will restrain inflation.5
Warsh, in turn, argues that despite the U.S. economy running at full employment and with inflation above the Fed’s target level of 2 percent, lower interest rates will not lead to higher inflation because of the acceleration of AI technology. In Warsh’s view, the development of AI technologies will rapidly increase economic output through increased productivity, allowing the Fed to cut interest rates aggressively without reigniting inflation.6
The purpose of this issue brief is to take these economic arguments at face value and see whether they are supported by standard macroeconomic theory or by the available evidence. To preview the conclusion, these statements by President Trump and Warsh are either inconsistent with standard theories or unsupported by factual evidence. A more cynical explanation is that the president’s interest rate demands are politically motivated and that Warsh’s nomination is meant to increase presidential influence over the Federal Reserve.
This is historically dangerous economic terrain. President Trump’s politically motivated attempts to influence the Federal Reserve by weakening its independence falls foul of abundant evidence of the benefits of central bank independence for robust macroeconomic performance and of the significant costs, in terms of inflation and economic output, of past episodes of political pressure on monetary policy.
Let’s begin this issue brief by examining President Trump’s economic reasoning.
President Trump’s statements, taken as a whole, seem to indicate that he does not believe in the existence of any trade-off between inflation and economic output. In contrast, modern macroeconomic theory, well-supported by empirical data, puts this trade-off at the heart of the aggregate supply relationship, wherein high economic activity that approaches or exceeds the capacity to supply goods and services will ultimately raise the demand for workers, oil, and other inputs, which will, in turn, raise prices.
The negative relationship between inflation and unemployment is known as the Phillips curve. An extensive economic literature has studied the slope of the curve, showing that the trade-off exists,7 even if its magnitude is not always clear and changes over time. Recent economic literature suggests that the slope is particularly large when unemployment is low.8 That means additional demand is more likely to translate into inflation when the labor market is already tight.
This well-understood relationship helps explain why inflation surged after the COVID-19 pandemic in 2020. Labor markets were particularly tight, and there were significant supply shocks from the war in Ukraine and other supply-chain disruptions that resulted in inflation rates close to double digits in 2021–2022. Today, the latest jobs report for April 2026 finds that the U.S. unemployment rate held steady at 4.3 percent, closely in line with Federal Reserve projections of the long-run rate of unemployment, suggesting that the inflation costs of higher demand are elevated. In this environment, excessive rate cuts could bring back higher inflation.
The inflationary costs of lower interest rates also are determined by the extent to which expectations of future inflation among consumers and businesses are anchored—that is, to what extent they believe the Fed will be able to achieve its inflation target of 2 percent.9 During the Great Recession of 2007–2009, well-anchored expectations from 20 years of inflation-targeting success led to fairly small deviations in inflation, despite a massive drop in demand.
In contrast, today, the U.S. and world economies are coming off the largest spike in inflation since the 1980s, a period in which household inflationary expectations became more dispersed and less stable.10 Another bout of elevated inflation would further reduce the credibility of the central bank’s ability to control price growth, leading to unanchored expectations and spiraling rounds of inflation.
Should President Trump succeed in lowering interest rates and spurring demand, any higher economic output and lower unemployment would be short-lived. That’s because the second form of the Phillips Curve, the “long-run Phillips Curve,” states that the trade-off between inflation and economic output is only available in the short run.11 Over longer periods, monetary policies that seek to permanently lower the unemployment rate beyond its natural limit will ultimately fail.
The reason for the failure to boost long-run output is that, over time, high inflation will affect the beliefs and expectations of consumers and firms. They will start to bake it into their prices and wage demands. Sustained higher expectations will then cycle back into higher prices, confirming the expectations in a vicious cycle. This was the experience of the U.S. economy in the 1970s and 1980s.12 In the long run, the Phillips Curve is vertical, meaning that there is no effect of inflation on output but rather that the economy ends up with higher prices but not more real economic activity.
Crucial to the Phillips Curve theory is the notion that unemployment cannot be lowered past its sustainable level, known as the natural rate of unemployment. Estimates of this natural rate are noisy, and they can change over time.
History teaches that inaccurately estimating the natural rate of unemployment has serious consequences. In the 1960s, Federal Reserve forecasts of the natural rate of unemployment were too optimistic, which led to unsustainably loose monetary policy at the time.13 In contrast, during the 1990s, the Fed’s forecast for natural unemployment may have been too pessimistic, with unemployment consistently low and modest inflation. This is why it is important today for Federal Reserve officials to independently estimate the natural rate of unemployment without undue political pressure.
President Trump also argues that supply-side policies can allow the U.S. economy to grow without higher inflation and therefore that interest rates can be safely cut to support this growth. Standard macroeconomic theory teaches that a decrease in the cost of resources, by increasing aggregate supply, can indeed raise economic output without putting upward pressure on inflation.14 The implications for the conduct of monetary policy, however, do not straightforwardly call for a decline in interest rates. This is because the appropriate interest rate depends not only on the disinflationary effect of the productivity shock, but also on how persistently that shock raises supply and thus affects the natural real interest rate.
In other words, managing the dynamics of AI innovation requires a calculated decision by a central bank that is carefully taking the pulse of both inflation and output growth.
The two supply-side policies that President Trump emphasizes the most are higher tariffs and the deregulation of the oil and gas industry.15 The logic that such policies will spur growth without raising inflation is predicated on the argument that they will reduce prices. On the contrary, the latest empirical research shows that a direct result of his recent tariffs has been an increase in prices, contributing to a 0.7 percent higher inflation rate.16 And while oil and gas prices had moderated early in his second term, they have risen dramatically due to the ongoing war with Iran, with global oil prices at more than $100 a barrel.
Contrary to President Trump’s inflation optimism, the most recent inflation reports show inflation above the Fed’s target of 2 percent. The April 2026 Consumer Price Index report showed 3.8 percent year-over-year inflation,17 and the March Personal Consumption Expenditures report showed 3.5 percent inflation.18 As of May 2026, the data do not contain evidence of a supply-side economic expansion.
The new Fed Chair Kevin Warsh has offered a different supply-side theory for lower interest rates, centered on the acceleration of artificial intelligence, which he says is “the most productivity enhancing wave of our lifetimes—past, present and future.”19 In Warsh’s argument, the development of AI technologies will rapidly increase output, allowing the Fed to cut rates aggressively without reigniting inflation.
This reasoning contains a grain of truth: If the promise of AI productivity pays off, there will indeed be reduced inflationary pressure. But that conclusion is premature. General-purpose technologies often take years or decades to diffuse throughout the economy.
So far, aggregate productivity data do not show a dramatic break from recent experience. To paraphrase the late Nobel laureate Robert Solow’s remarks on computer technology, the effects of AI are everywhere except in the data. In the latest data from the U.S. Bureau of Labor Statistics, year-on-year total factor productivity growth was just 0.8 percent, while labor productivity growth was 2.5 percent. There is no indication in the data of an explosion of productivity that would justify lower interest rates.
President Trump’s theories do not accord with standard economic theory, but they do fit a long historical pattern of presidential pressure on the Federal Reserve to lower interest rates at the expense of long-term monetary stability. This problem of political pressure motivated the introduction of independent central banks, which are not under the direct purview of executive or legislative branches.20 Independence gives central banks an arm’s-length distance from the executive and legislative branches. These branches set the central bank’s monetary policy objectives but ideally do not interfere directly with the operational conduct of monetary policy itself.21
Historically, central bank independence has been honored more in the breach than in the observance. Presidents prefer lower interest rates, and they lean on their central bank chairs to achieve their political goals. Noted violations of Fed independence include President Richard Nixon’s pressure on Fed Chair Arthur Burns, as well as President Lyndon B. Johnson’s pressure on Fed Chair William McChesney Martin. This political pressure has been shown to empirically increase inflation with no impact on economic growth.22
President Trump has repeatedly violated the principle of Fed independence by publicly demanding lower interest rates and initiating politically motivated investigations of Powell and Fed governor Lisa Cook. These actions can only lead to a further deterioration of the Federal Reserve’s credibility, with deleterious consequences for long-term inflation and growth.
Current deregulatory policies under the second Trump administration are not leading to an increase in the supply of basic economic inputs. Tariffs are not spurring a supply-side economic expansion. And while AI has the potential to drive a new wave of economic growth, it is not yet showing up in significant increases in productivity.
Meanwhile inflation remains elevated, in part due to the very policies intended to lower it—tariffs—and to unrelated but no less significant policy choices, such as military action against Iran and the consequent rise in global energy prices and the cost of fossil-fuel based inputs such as fertilizer.
Lower interest rates are politically popular because they can temporarily juice the economy. History demonstrates that the incumbent party’s chances of electoral success are higher when the economy is strong. Yet history also shows that loose monetary policy not anchored to underlying economic conditions can have deep and long-term negative effects on the economy, stunting growth and stoking stubbornly high inflation.
Such a wave of stagflation would have real and significant costs for workers and families, making daily life more expensive while decreasing the availability of jobs and therefore depressing wages. It is that reality that motivated the United States and the rest of the developed world to enshrine the independence of central banks and insulate monetary policy from political pressure.
President Trump and his new Fed Chair Warsh should heed these warnings.
Essays SEP 8, 2026
By: Alexander Bick, Adam Blandin, David Deming, Tyler Schumacher
Working Papers SEP 8, 2026
By: Alexander Bick, Adam Blandin, David Deming, Tyler Schumacher
Essays SEP 2, 2026
By: Megan Rivera
Get updates on our latest research, event announcements, and policy insights delivered straight to your inbox. Stay connected with the leading voices on equitable growth.