For years, central banks viewed artificial intelligence as they did climate change—a distant, structural trend to be monitored from afar, but not a force that directly shaped their monthly decisions on interest rates. That era of detached observation is decisively over. Policymakers now openly frame AI as a historical pivot on par with electrification or the internet, a transformation that will reconfigure inflation dynamics, long-term growth, and the very models used to steer the economy. The core debate has shifted from whether AI matters to pressing questions of timing, transmission, and direction. The critical unknown is whether this technological wave will first unleash a disinflationary surge of productivity or ignite a near-term inflationary investment boom—and how monetary authorities should navigate these opposing forces.
Leading the operational charge, the European Central Bank and Germany’s Bundesbank have already integrated AI into their daily work. In a landmark disclosure, ECB economists revealed that a machine learning model has been part of the toolkit for preparing monetary policy decisions since late 2022. This system analyzes dozens of real-time indicators—from inflation expectations to financial conditions—and has already proved its worth, successfully flagging upside risks to core inflation ahead of time in 2025. Across the Rhine, Bundesbank President Joachim Nagel has detailed a suite of AI applications in use, from intelligent assistants to a model named MILA that analyzes communications from other central banks. For these institutions, AI is no longer theoretical; it is a practical tool to “fulfil our mandate as well as possible,” as Nagel stated, enhancing human analysis with unprecedented speed and pattern recognition.
At the U.S. Federal Reserve, the integration is more conceptual but no less urgent, with officials vigorously debating how AI reshapes the fundamental trade-offs of monetary policy. Governor Christopher Waller has placed the productivity question at the very center of the policy debate, noting AI’s rapid adoption and expressing hope that it will deliver sustained growth without inflationary pressure. Vice Chair Philip Jefferson, however, highlighted the double-edged nature of the technology, pointing out that while AI may lower costs through efficiency, its massive infrastructure demands—for data centers, energy, and land—could push input prices higher. The debate took a politically charged turn with Fed nominee Kevin Warsh, who likened the AI boom to the productivity miracle of the late 1990s but warned it is approaching “escape velocity.” He cautioned that while AI may ultimately ease price pressures, policymakers cannot yet bank on those gains, especially given uncertainties around its impact on employment, the other half of the Fed’s dual mandate.
This divergence in interpretation has split financial markets into two opposing camps. On one side are the “disinflation bulls,” who treat AI as a monumental positive supply shock. Figures like Mike Hunstad of Northern Trust argue that AI-driven productivity could accomplish the disinflationary work that years of high interest rates could not finish, potentially leading to lower prices, lower long-term rates, and higher valuations for risk assets. In stark opposition stand the “capex hawks,” who foresee a near-term inflation problem. They point to an unprecedented capital expenditure cycle, where booming investment in AI infrastructure strains power grids and competes for resources. As noted by Oxford Economics and Goldman Sachs, this is already elevating electricity prices—a significant and underappreciated inflation channel—and could drain savings, lift bond yields, and boost aggregate demand long before any efficiency gains materialize to offset them.
The central question, therefore, is one of sequence, not ultimate destination. There is a growing consensus that AI’s economic impact will be profound, compelling a fundamental rethink of economic models. Yet the path remains fraught with uncertainty. If meaningful productivity enhancements arrive swiftly, they could give central banks the coveted “goldilocks” scenario: room to ease policy to support growth without rekindling inflation. Conversely, if the massive investment phase hits the economy first—driving up energy costs, capital demand, and asset prices—policymakers who cut rates preemptively on the promise of future disinflation could be forced into a painful and credibility-damaging reversal. This delicate balancing act marks a radical departure from the recent past, where AI was a mere footnote in monetary policy discourse.
We now stand at an inflection point. Artificial intelligence has collapsed the distance between a distant technological future and the immediate concerns of monetary policy. Central banks are no longer mere observers; they are becoming active users of the technology while wrestling with its macroeconomic consequences. Their challenge is to navigate a landscape where the same force promises both inflationary headwinds and disinflationary tailwinds, separated only by time. The decisions made in this period of transition—whether to wait for proof of productivity or act on its promise—will test the judgment of policymakers and ultimately shape the economic legacy of the AI age. The long horizon has arrived at the meeting room door.
