The economy is already struggling with inflation and high living costs. What happens when AI starts removing millions of paychecks from it?

I don’t think people are thinking through the chain reaction here. AI doesn’t have to cause a recession by itself. It could make an existing economic problem much worse.

Look at the timing. The October jobs report showed 165,000 new jobs and unemployment at 4.1%. Sounds pretty good. But the long-term unemployment share climbed to 27.6%, and information and finance remain among the weakest industries. Goldman Sachs recently estimated that up to 9% of workers could be displaced during the AI transition, with slower hiring putting another 3% pressure on employment. That’s potentially a huge number of people competing for fewer jobs.

And where exactly are those people supposed to go? Someone who spent 15 years in accounting, finance, marketing, software development or customer service can’t just take a six-week course and walk into another career. Retraining takes time. Employers need fewer people in the first place. Meanwhile, AI keeps getting better and cheaper. The cost barrier is coming down, and companies don’t need to be losing money to replace expensive employees with software. They just need to see a better return.

We’re already seeing companies make the cuts while redirecting money toward AI. Meta announced 8,000 layoffs and plans to leave another 6,000 positions unfilled. Block is eliminating more than 4,000 jobs from a workforce of roughly 10,000. Coinbase is cutting 14% of its staff. Barclays plans to eliminate 5,000 positions. Not every one of these cuts is purely because of AI, but the direction of corporate spending is pretty clear.

Then there’s the consumer side. People lose jobs or worry about losing them, they spend less. Businesses lose revenue, cut more costs, and invest even more in automation. The people who used to buy the products and services are now the people being replaced. That’s a problem you can’t solve just by telling everyone to learn AI.

Today’s inflation report showed consumer prices rose another 0.4% in October, following a 0.6% increase in September. Energy prices rose 0.7% in October after jumping 5.2% the previous month. Real average hourly earnings were unchanged. People are already struggling to keep up with prices. Now imagine adding a wave of white-collar unemployment to that.

There’s another contradiction in the AI boom. Token prices have fallen 54% from their May peak, making AI cheaper to use. At the same time, the largest U.S. hyperscalers are on track to spend around $800 billion on capital expenditures this year, with another $1.1 trillion projected for 2027. AI usage keeps climbing, but business spending on tokens was flat for two consecutive months, according to Ramp data cited by Axios. Companies are spending enormous amounts to build the infrastructure while trying to spend less on the people who work for them.

And the government has very little room to absorb a serious downturn. The federal deficit reached roughly $2 trillion in fiscal 2026. Interest expense on debt held by the public hit about $1.1 trillion. Federal debt subject to the debt limit passed $40 trillion in October. If unemployment rises sharply, demand for assistance increases and tax revenue takes a hit. Meanwhile, the government is already paying enormous sums just to service its debt.

This is why I think taxing extreme wealth needs to be part of the discussion. The people and companies benefiting most from AI-driven productivity gains shouldn’t be the only ones collecting the rewards while everyone else absorbs the disruption. Revenue could help fund social safety nets, retraining and public services. Tighter AI regulations also need to be considered before companies automate entire departments simply because the technology makes it financially attractive.

The problem is that technological progress doesn’t wait for displaced workers to find another job. And the economy depends on those workers continuing to earn money and buy things.

We keep discussing AI productivity gains. I’m more interested in what happens to the economy when a growing number of people no longer have a paycheck to spend.