Understanding the Fed's AI doom loop

A version of this article originally appeared in Quartz’s members-only Markets newsletter. Quartz members get access to exclusive newsletters and more. Sign up here.

Here’s a big, provocative question: What if the Federal Reserve is funding the very problem it’s trying to solve? More precisely: What if the Fed’s current round of interest rate cuts, which are explicitly intended to support a struggling jobs market, ends up accelerating the very technology that’s contributing to job losses?

I raise this question because so many massive economic forces are colliding at once, in front of our eyes, and the territory is arguably uncharted. More data emerged last week suggesting that the labor market is weakening. At the same time, the headlines about how AI is reshaping white-collar labor demand keep coming, even as some of the world’s most famous bears call attention to AI’s debt and finance dynamics. Such forces seem esoteric until you consider that, if you’re reading this, you’re likely a white-collar worker with a stock portfolio, a property portfolio, or both. 

But to understand how these pieces interact — and why the Fed’s decisions could have unintended consequences — we need to start by laying out the key facts. 

First of all: What job losses? 

According to last week’s ADP report, private payrolls once again “unexpectedly” contracted in November — a jarring signal that the labor market is weakening in the exact places policymakers least want to see.

Employers cut 32,000 jobs, with the losses concentrated in white-collar industries. Professional services shed 26,000 positions, information declined by 20,000, and finance fell by 9,000. ADP’s data now show private payrolls have been effectively flat since July, quite a lengthy stall for a supposedly “cooling yet resilient” economy.

Comerica’s chief economist framed the report bluntly, saying the U.S. appears to have hit an economic “air pocket.” Small businesses seem to be suffering the most — firms with fewer than 50 employees cut an eye-watering 120,000 jobs in November — while gains are increasingly confined to healthcare and other nondiscretionary sectors. 

With official BLS reports still delayed by the government shutdown, the Fed is likely to treat ADP’s data as a more central input than usual heading into this week’s policy meeting.

What rate cuts?

Before the ADP release, prediction markets were already leaning heavily toward a rate cut. Early last week, Kalshi traders put the odds of a 25-basis-point cut at roughly 88%. Then the ADP report hit and the probability jumped to 93%. 

So it’s more than reasonable to guess that Fed Chair Jerome Powell will announce another rate cut this week, and that labor-market weakness will be central to the FOMC’s reasoning. Powell already cited the labor market when cutting rates in September and October. Now it looks like we’re cruising in for a hat trick. 

There’s significant historical precedent for Powell’s thinking. Rate cuts have long been the Fed’s most reliable tool for supporting the labor market. The logic is straightforward: When the Fed lowers interest rates, borrowing becomes cheaper for households and businesses. Cheaper credit encourages consumers to spend more and firms to invest more, fueling demand across the board. As economic activity picks up, employers hire additional workers to meet that demand. 

This mechanism describes basically all post-war macroeconomic management. And Powell operates squarely within the tradition, essentially arguing that, while the Fed can’t create jobs directly, it can help engineer the financial conditions in which job creation naturally accelerates. 

And far be it for me to say Powell is wrong or mistaken. In normal times, monetary easing works reliably, if slowly. Across administrations, Fed chairs, and economic cycles, the belief in the connection between rate cuts and labor-market strength has been remarkably consistent — and remarkably successful.

So what’s different today? 

It comes down to AI. Rapid technological shifts aren’t new. But whether AI works in the same way as previous shifts is far from a settled question. Is the current bout of labor market weakness cyclical or structural? 

A growing body of evidence points to the latter, suggesting that growing labor-market weakness is structural. For instance, a recent MIT study estimates that roughly 12% of the U.S. workforce could already be replaced by existing AI technology — not hypothetically in the future, but right now, with present-day models. 

The MIT figure isn’t a forecast of coming layoffs. Instead, it reflects the share of U.S. workers whose tasks can already be executed by modern AI systems. That’s one in nine American workers, representing $1.2 trillion in wages. MIT distinguishes between “visible” job losses — like tech layoffs, concentrated in tech cities — and “hidden” displacement in fields like HR, logistics, finance, and administrative work, which are spread across all 50 states. However you look at it, the finding is stark: A large slice of the U.S. labor market is newly exposed to automation in ways the larger economic cycle can’t and doesn’t explain. 

As an example, look at Amazon. Long one of the country’s largest employers, Amazon is experimenting aggressively with robotics and AI systems that can perform significant portions of warehouse and logistics work.

Officially, Amazon insists these systems are meant to “extend human capacity,” not replace it. But internal memos reported by the New York Times suggest long-term scenario planning that would see 75% of operations automated, eliminating the need to hire for hundreds of thousands of future jobs. Even if existing Amazon employees aren’t cut en masse, the hiring baseline shifts downward, meaning vastly fewer entry points for new workers over time. 

Meanwhile Amazon has already announced significant white-collar layoffs — about 14,000 in recent months — with executives explicitly tying some reductions to internal productivity gains from AI tools. So you can see the automation trends hitting both blue-collar and white-collar workers, spanning warehouses and corporate offices. 

In a nutshell? You could argue that a significant portion of today’s labor-market weakness is the result of firms across the country beginning to discover that they can perform large portions of their businesses with dramatically fewer humans. That may look roughly like the results of a broad-based downturn. But it’s meaningfully different, because it’s not actually driven by downturn dynamics. 

And that matters, because a response that loosens credit conditions could accelerate the structural labor weakness, rather than shoring it up. How exactly? As complex as the moving parts may be, the answer is simple: If the Fed cuts rates and makes borrowing cheaper, it could make it easier for more companies to make the very speculative AI bets driving job displacement. 

A closer look at the credit piece

We already know that developing and deploying AI systems is capital intensive, period. It takes money to devise the tech, and it costs money for enterprises to deploy it inside their businesses.

Lower rates make AI infrastructure financing cheaper. That matters less for the Amazons of the world, with their unequivocal prime borrower status and huge cash flows, and much more for smaller, less-economic players who are already more likely to take wild swings. Companies are already borrowing billions for data-center buildouts; cheaper debt accelerates those projects. Those projects (arguably) accelerate job-loss trends. 

And for individual companies investing in AI to streamline their businesses and effectively lower or limit their headcounts, lower rates improve the ROI calculation on automation investments. When you can borrow at 5% instead of 6%, replacing a $60,000 employee with a $40,000 software subscription pencils out faster. In short, easier credit conditions could mean more mid-sized companies can afford enterprise AI tools they couldn't justify before. 

Meanwhile, lower rates also tend to boost equity valuations for tech companies, giving them yet more financial firepower for AI acquisitions and internal development. 

Viewed in this light, there’s a case for seeing the Fed as essentially subsidizing the AI boom that's reshaping the labor market.

OK, so do we KNOW this is happening?

The short answer is no. 

Whether this dynamic is already playing out or will become more visible over time remains uncertain. But the mechanism is real: Cheaper borrowing costs lower the bar for automation investments. If the Fed continues cutting rates while AI-driven displacement accelerates, we'll have a clearer answer about whether monetary policy intended to support employment could be inadvertently financing its transformation.  It may be a dark thought exercise, but it’s also a useful one that reframes some of the most important economic issues of our time.

Which, if you’ve been reading Quartz Markets, you know is the sort of thought exercise we enjoy around here — and hope you enjoy, too.

To be clear, Powell doesn't have perfect options. If he holds rates steady, any cyclical weakness could spiral into outright recession  — far worse for workers. If he cuts, he stimulates demand and supports employment through traditional channels, but potentially accelerates automation through cheaper capital. 

Either way, monetary policy arguably now comes with side effects that previous Fed chairs didn't have to consider, namely the potential to make it cheaper to replace workers just as the technology to do so becomes viable. The risk is not that the Fed is wrong. It’s Murphy’s Law, because policy tools now run through and across technologies that affect labor demand in new and difficult-to-manage ways. That’s the thought to take with you into this week, when we’ll get that all-important rate-cut news.