Fed's Barr Names 3 AI Scenarios — Including 'Essentially Unemployable'

A US Federal Reserve governor used the phrase "essentially unemployable" out loud last month — and he wasn't talking about a fringe scenario. He was talking about one of three possible futures the Fed is now actively planning around.

That phrase came from Michael S. Barr, the Fed's Vice Chair for Supervision, in a February 17, 2026 speech at the New York Association for Business Economics. The speech is the most direct, structured Fed commentary on AI's labor market impact to date. And while the headlines focused on the scary scenario, the more interesting story is the one Barr says we're _currently_ closest to — and the one he says could blow up the rate-cut narrative. [Fact]

If you have been wondering what central bankers actually think about AI taking jobs, the answer is now on the record. Here's what Barr laid out, what data he leaned on, and what it means for workers.

The three scenarios the Fed is planning around

Barr does not project a single AI future. He sketches three, and explicitly says current data is "closest" to the first. [Fact]

Scenario 1 — Gradual adoption. AI diffuses like prior general-purpose technologies, perhaps a bit faster. You get strong productivity growth comparable to the late 1990s and early 2000s. Some occupational displacement happens, but new roles emerge. The pace is slow enough that workers have time to retrain, and real wages rise with productivity. This is the base case.

Scenario 2 — Rapid growth, the "jobless boom." AI capabilities grow exponentially. Adoption is extremely fast. Professional and service occupations get displaced, autonomous vehicles and robotics eliminate manufacturing and transportation jobs, and AI-centric startups with radically new business models replace incumbent firms. Barr's exact words: "widespread unemployment in the short run and declines in labor force participation over time, as a large share of the population is essentially unemployable." [Fact] He treats this as plausible but not the central forecast.

Scenario 3 — Stalled growth, the "AI bust." Training data runs out. Electricity supply for data centers becomes a binding constraint. Capital dries up. AI delivers modest gains on easy tasks, fails on complex ones, and ends up like email or search engines — ubiquitous but not revolutionary. The risk in this scenario shifts from labor to financial markets, with possible bond defaults reminiscent of the dot-com bust. [Fact]

The honesty here is unusual for a Fed speech. Most central bankers pick a forecast and stick to it. Barr explicitly says the future is uncertain across an order of magnitude, and the Fed's job is to be ready for any of the three.

How much AI is actually being used right now

Before you panic about the jobless boom, look at the adoption numbers Barr cited. They are smaller than the discourse suggests.

According to the December 2025 US Census Business Trends and Outlook Survey, 17% of US firms reported using AI in business functions. [Fact] Among large firms with 250 or more employees, the share rises to about 30%. [Fact] A McKinsey survey of mostly large firms is much higher — 88% report AI use in at least one business function — but McKinsey's sample skews toward the firms most likely to be early adopters. [Fact]

Generative AI specifically went from 33% adoption in 2023 to 79% in 2025 within McKinsey's sample. [Fact] That is fast. But even Barr noted the pace is comparable to IBM PC adoption after 1984 — which felt revolutionary at the time and still took roughly 20 years to fully reshape office work.

In other words: the technology is real, the adoption is accelerating, but most US firms have not yet integrated AI deeply enough for it to show up in workforce statistics. Yet.

The early-career signal that nobody is talking about loudly enough

This is the part of Barr's speech that should worry workers more than the "essentially unemployable" headline.

He cited research showing that early-career workers in AI-exposed occupations — software developers and customer service representatives among them — have already experienced a decline in employment relative to other early-career workers. [Fact] Aggregate employment effects are small because firms are shifting workers to complementary tasks. But the entry-level door is narrowing in specific fields.

This matters because of a well-documented economic pattern: entering a weak labor market early in your career has persistently adverse effects on your lifetime earnings. The cohort that graduates into the AI-exposure squeeze is at risk of carrying the cost forward for decades, even if aggregate employment looks fine.

If you are advising someone choosing a career path, this is the piece of the data to take seriously. The aggregate "employment is stable" story does not apply uniformly across age groups.

What Barr says about productivity — and the rate-cut narrative

Here is the part that should interest anyone watching interest rates.

Barr cited research from Filippucci et al. (2025) projecting that AI could contribute 0.3 to 0.9 percentage points to annual total factor productivity growth over the next decade. [Fact] The upper end of that range matches the strong-growth period of the late 1990s when internet communications technology drove a productivity boom. The Federal Reserve's own internal example: one database migration project where AI tools cut completion time by 50% and detected 30% more issues during testing. [Fact]

Why does this matter for rates?

Because if AI genuinely raises productivity, it pushes up the long-run neutral rate of interest — what economists call r-star. Higher productivity means stronger business investment (capital demand goes up). Higher expected lifetime earnings mean lower household savings (capital supply goes down). Both push the equilibrium interest rate higher. Barr said directly: "I expect that the AI boom is unlikely to be a reason for lowering policy rates." [Fact]

If you have been waiting for the Fed to cut rates aggressively, the AI productivity story is working _against_ that hope, not for it. Add in the inflationary pressure from electricity demand at AI data centers, and you get a Fed that may stay tighter for longer than markets are pricing in. [Opinion]

The monetary policy bind

Barr was unusually clear about the limit of his own institution's tools. He said monetary policy "cannot address the structural factors that determine long-run employment." [Fact] If AI displaces workers structurally — not cyclically — rate cuts will not bring those jobs back. The right tools are education, training, and workforce policy, none of which the Fed controls.

This matters because it changes who is responsible for the AI labor transition. The Fed can stabilize aggregate demand. It cannot retrain a 45-year-old paralegal whose document review work has been automated. That responsibility sits with Congress, state governments, and employers — and the historical record on workforce transition programs, in Barr's words, "is not encouraging." [Fact]

If you are in a high-exposure occupation — and the Yale Budget Lab data suggests that includes many data scientists, financial analysts, accountants, administrative assistants, and computer programmers — the implication is uncomfortable. Even if displacement happens at scale, the policy infrastructure to catch you is weak. Plan accordingly.

What this changes about how to read AI labor news

Three takeaways from Barr's speech that should change how you read everything else.

First, the Fed is no longer treating AI as a niche topic. A Vice Chair gave an 18-page formal speech on it. That is the institutional version of declaring something is now first-order important.

Second, the base case is gradual, not catastrophic, but the entry-level signal is real. Aggregate stability and early-career displacement can coexist — and the second matters enormously for individual career decisions even if the first dominates the headlines.

Third, the rate cut you have been hoping for might not come. If AI boosts productivity meaningfully and lifts the neutral rate, the Fed will need to keep policy tighter than markets currently expect. That has implications for everything from mortgage rates to startup funding to your 401(k).

Sources

• Federal Reserve Vice Chair for Supervision Michael S. Barr, "What will artificial intelligence mean for the labor market and the economy?", New York Association for Business Economics, February 17, 2026 (republished by BIS Review)
• Filippucci, F., et al. (2025). Cited by Barr for AI productivity projection of 0.3-0.9 percentage points to annual TFP growth.
• US Census Bureau, Business Trends and Outlook Survey, December 2025. AI adoption statistics for US firms.
• McKinsey Global Survey on AI, 2025. Generative AI adoption trajectory.

---

_This article was written with AI assistance and edited for accuracy. Statistics are current as of the source publication dates. Quotations from Vice Chair Barr's speech are taken from the BIS Review republication._