Can the Fed keep up with the AI economy?

The U.S. Federal Reserve has always adjusted to technological change—eventually. But the arrival of generative AI and advanced automation isn’t just another step forward in productivity. It’s a potential leap. Tools like ChatGPT, GitHub Copilot, and enterprise AI assistants are reshaping how work is done, slashing time costs and redistributing tasks across industries. Yet, despite the pace of change, the Fed’s policy framework remains anchored in a 20th-century understanding of inflation, labor markets, and output gaps.

There’s a growing disconnect between how fast AI is transforming the economy and how slowly institutions like the Fed are adapting their models. If policymakers fail to respond with new tools, measures, and mental models, they may misread the signals and misjudge the stakes.

One of the clearest promises of AI is productivity growth. Goldman Sachs estimates that AI could raise global GDP by 7% over the next 10 years. McKinsey argues that generative AI could add trillions of dollars in annual economic value. But this optimism masks a familiar paradox: the data doesn't yet show the boom.

In the 1990s and early 2000s, it took years for digital technologies to show up in national productivity metrics. We’re now seeing early signs that AI might follow the same path—or present even deeper measurement challenges. For example, a legal firm might replace paralegal review tasks with a chatbot, reducing labor hours dramatically. But GDP might only register a drop in employment, not a surge in output. The same goes for a customer support team that’s quietly replaced by an LLM.

This lag matters because the Fed depends heavily on backward-looking indicators to shape forward-looking policy. If real productivity is accelerating under the surface—but official data still suggests stagnation—the Fed may misread the situation and hold interest rates higher for longer, mistaking supply-side improvements for stubborn demand.

One of the most profound disruptions AI introduces is in the labor market. And here, too, the Fed’s models may be out of sync. Today’s employment reports may show low overall unemployment, but that masks the rapid disappearance of entire job categories. From legal assistants to customer service reps, entry-level marketing analysts to junior programmers, automation is eating away at white-collar roles.

Meanwhile, hiring is concentrated in specialized roles—prompt engineers, data scientists, AI ethicists—requiring highly specific training. The result? A bifurcated labor market where one segment sees rising wages and high demand, while another faces stagnation or displacement.

The Fed’s dual mandate—maximum employment and price stability—was designed in an era where labor was a broad and reliable inflation signal. But in an AI economy, full employment may no longer mean broad employment. If productivity surges due to automation, inflation may stay low even as millions struggle to re-enter the workforce. This decoupling risks undercutting the Fed’s ability to use employment as a reliable economic indicator.

Artificial intelligence also complicates how we think about inflation. Many AI tools are embedded in services that are either free or priced differently from traditional goods. Consider a customer support chatbot that handles thousands of inquiries at near-zero marginal cost. Or an AI design tool that does the work of three junior staffers overnight. These innovations drive deflationary pressures—but they’re hard to quantify using the Consumer Price Index (CPI).

The Bureau of Labor Statistics doesn’t yet have the capacity to price software productivity in real time. So while AI adoption may be slashing operating costs, reducing time-to-market, and boosting consumer utility, those gains may not show up in headline inflation. Worse, they may create a false narrative of “sticky prices” if traditional sectors are slower to pass on cost savings.

In this way, the Fed may be tightening policy to fight inflation that doesn’t reflect the true state of the economy—ignoring embedded deflation in the most dynamic corners of the digital economy.

The Fed’s traditional toolkit—manipulating interest rates to slow or stimulate the economy—was built for an industrial world. But what happens when monetary transmission weakens in an AI-powered economy? For example, a rate hike that curbs consumer spending may have little effect on enterprise decisions to roll out AI systems. Software-as-a-service models, with low upfront costs and rapid returns, may be less sensitive to financing conditions than factory expansions or home mortgages.

Moreover, if companies are replacing labor with AI and boosting output per worker, the usual wage-driven inflation loop may weaken. Raising rates to slow wage growth may hit the wrong targets: workers who are already being left behind. There’s also the risk of overshooting. If the Fed raises rates too aggressively in a bid to control inflation that’s actually being softened by AI-driven productivity, it may choke off investment and delay the broader diffusion of AI benefits.

There’s another layer to consider: investor psychology. The Fed’s language and posture play a huge role in shaping market expectations. If the central bank appears overly cautious about AI’s effects, markets might interpret this as either denial—or quiet endorsement of the hype. The result could be a misalignment between macro fundamentals and asset prices.

Tech stocks, particularly those tied to AI infrastructure, have soared in anticipation of a so-called “AI supercycle.” Nvidia, for example, recently hit a record valuation, reflecting investor bets that AI will drive demand for years. But if the Fed is misreading the nature of these shifts—seeing them as speculative bubbles rather than capital deepening—it may misjudge systemic risk.

This is not purely hypothetical. Missteps in interpreting dot-com valuations or housing price inflation have led to policy inertia in the past. If AI becomes another blind spot, the consequences could ripple through credit markets, labor expectations, and national growth trajectories.

Implications:

1. Update Models, Fast
The Fed’s existing DSGE (dynamic stochastic general equilibrium) models struggle to incorporate rapid technological change. AI forces a rethinking of productivity assumptions, labor substitution effects, and inflation sensitivity. It’s time to revise these models—not just tweak them.

2. Build Real-Time Economic Indicators
Lagging data on productivity, wages, and inflation won’t cut it in an AI-driven economy. The Fed should invest in real-time economic dashboards that integrate private-sector datasets—from software usage rates to job platform analytics—to get a timelier picture of change.

3. Rethink the Natural Rate of Unemployment
If AI compresses the gap between output and labor, the Fed’s estimate of the NAIRU (non-accelerating inflation rate of unemployment) may be too high. This would imply a structural shift in labor dynamics that requires new policy thresholds.

4. Coordinate with Fiscal and Regulatory Policy
AI is not just a monetary issue—it intersects with education, retraining, antitrust, and digital infrastructure. The Fed can’t steer this alone. It needs to coordinate with other agencies to ensure the broader economic scaffolding is ready for disruption.

The Federal Reserve is facing a new kind of economy—one where software outpaces statisticians and algorithms quietly displace entire categories of labor. This isn’t just a policy challenge; it’s an existential one. If the Fed continues operating under assumptions from the pre-AI era, it risks becoming irrelevant—or worse, reactive.

We believe the Fed needs to pivot fast. That means integrating AI literacy into its forecasting teams, reengineering its inflation models to include digital deflation, and developing tools for scenario planning in a rapidly shifting labor landscape. It also means being honest: about uncertainty, about blind spots, and about the limits of monetary policy in shaping a future built by machines. Artificial intelligence won’t wait for institutional reform. If the Fed doesn’t adapt, it won’t just miss the signal. It might become the noise.