Who Captures the AI Dividend Will Decide How Long Inflation Lingers

Whose pocket the AI money lands in, and why central banks are alarmed first

On September 21, ECB Governing Council member Fabio Panetta told an audience in Kyiv that how the AI dividend gets distributed will directly reshape the direction of inflation.

Panetta also serves on the ECB Governing Council(the ECB's top decision-making body, made up of six Executive Board members and the governors of national central banks). Reuters reports he was speaking at an event hosted by the National Bank of Ukraine, where he put a usually sidestepped question at center stage: who ultimately captures the AI gains — workers, corporate shareholders, or a handful of tech giants?

He laid out two opposing paths. If AI creates new jobs first and lifts households' expected labor income, consumer demand swells before the productivity dividend materializes, prolonging price pressures. If AI takes the automation route instead, displacing labor and darkening wage prospects, households save more, and weaker consumption surfaces deflationary effects earlier.

Wages first or profits first: who calls the tune on prices?

Whose pocket fills up first decides where prices go. Productivity is half the story. Distribution is the other half.

Path one starts with wages. AI mainly creates new jobs and raises households' expectations of future income, so families feel flush and willing to spend. Demand picks up before the productivity dividend arrives, pushing prices upward and making inflation stickier.

Path two starts with profits. AI mainly replaces human labor, pushing down corporate costs and expanding output, but wages don't keep pace. Job and income uncertainty makes households clutch their wallets and save more, consumption weakens, and AI's price-suppressing effect arrives earlier.

The money hasn't been earned yet, but the market is already celebrating AI

AI companies themselves haven't posted decent profits, yet capital markets have already spent the money on their behalf. At the same event in Kyiv, Panetta raised a second warning: "current asset prices reflect optimistic expectations about AI's future profitability."

High valuations help tech companies raise the next funding round and make the next round of investments. But once those expectations fail to deliver, the market becomes extremely fragile. The phrase he used was short: a sharp correction.

Stocks aren't the only ones with the same affliction. Earlier this month, Bundesbank President Joachim Nagel called AI a "litmus test" for Europe, saying the continent cannot afford to fall behind. With the heads of the German, French, and Italian central banks speaking out in succession, this is no longer one economist's personal worry, but the eurozone's financial regulators jointly giving the market a warning shot. The party can go on, but the books still have to match reality.

The signals are already visible. Panetta noted that "recent volatility in tech stocks illustrates exactly this." The whipsaw in stock prices isn't because fundamentals changed; it's because the market's judgment of "how much money can AI actually make" is swinging wildly.

Will central banks dare to answer "whose pocket the money lands in"?

Panetta's verdict comes down to one line: central banks cannot stand on the sidelines. Credibility is every central bank's core asset, and how credibility is built and maintained depends on the nature of the challenges the central bank must face.

On September 21, Bank of Italy Governor and ECB Governing Council member Panetta went this far at an event hosted by the National Bank of Ukraine. Whether central banks step forward to answer "who gets the AI dividend" is itself shaping the market's trust in them.

Translated into operational language: central banks used to sketch out the inflation path by watching unemployment and wage growth. That empirical curve is no longer sufficient. The distribution of AI gains will simultaneously rewrite the consumption, investment, and wage curves, and central banks have to slot "the distribution channel" as a variable into their policy models, or predictions and reality will drift further apart.

He is speaking this forcefully because the cost of misjudgment has changed. The warning quoted earlier still stands: once expectations fail, the market is vulnerable to a sharp correction, and the recent volatility in tech stocks is a preview. A central bank that underestimates distributional effects could face two layers of pressure at once, with inflation going off course and markets crashing.

What to watch next? First, watch whether the next ECB and Federal Reserve policy statements and meeting minutes start to surface "distributional effects" or similar language. If they do, this causal chain has been formally written into the decision-making framework; if not, it remains in scholars' lecture notes.

Then watch the research pipeline. Will research papers and economic bulletins from Western central banks break out two separate AI benefit channels, "workers vs. capital," and try to quantify each one's pull on aggregate demand? Once comparable numbers are published, the distribution channel moves from concept to a monitorable indicator.

Finally, watch the stress tests. Panetta has already flagged the sharp swings in tech stocks as a leading indicator, but no central bank currently lists "AI expectations falling short" as a standalone scenario. If it happens, the risk has been claimed; if it doesn't, central banks are still using the old script to assess new risks.

Watch these five things until central banks spell out how the gains get split

Panetta didn't write a prescription. He named five things AI is about to rewrite: productivity, economic growth, labor markets, financial markets, and the payments system. Until central banks clarify how the gains get distributed, tracking those five is what readers can actually do.

Two of them are the most direct windows for ordinary readers: the labor market and financial markets. On the labor side, look at your own industry. Is AI creating new roles first, or trimming headcount first? The first maps to the inflation-stretches-out script, the second to the deflation-comes-early script. On the financial side, watch whether tech stocks stay near their highs. Panetta's warning is plain: asset prices have already pulled forward AI's future profitability, and once expectations fall short, the market will be fragile. Recent volatility is the signal.

On the central bank side, there is still very little they can do for now. Panetta's emphasis is that central banks cannot stand on the sidelines. They must understand the changes and dare to face the challenges, or credibility won't hold. Translated into actions readers can observe, it means watching two things: whether the ECB will, in its official documents, single out "AI gains distribution" as an independent policy variable for discussion, and whether the next ECB rate decision will include "distributional effects" or similar language. Neither has happened yet.

The payments-system thread is the one most interpretations skip, and Panetta singled it out anyway. It is also the cheapest place to look. Check whether the payment tools you use daily are starting to fold in AI for risk control or routing. If they are, "the reshaping of payments" is already underway — and you can see it without waiting for a policy document.

Productivity and growth are the slowest of the five to show up in anything you can personally observe. They surface in official statistics with a lag, which is exactly why the other three matter as early signals.