Goldman Sachs economists have been tracking a number that does not get enough attention. AI is erasing approximately 16,000 net jobs per month across the American economy. Not layoffs pending recall. Not positions temporarily eliminated. Net jobs gone, month after month, as AI systems take over the work that humans were doing.
The workers absorbing most of that damage are young, educated, and entry-level. Gen Z. The generation that was told a degree would protect them, that writing skills mattered, that getting into tech was the safe bet. All of those bets are losing simultaneously.
The Numbers Are Not Abstractions
The tech industry cut nearly 80,000 positions in the first quarter of 2026 alone. Almost half of those cuts were attributed directly to AI, by the companies doing the cutting. That is not speculation or projection. Those are the companies’ own explanations for why they eliminated the positions.
The roles disappearing fastest are the ones that used to serve as the entry point into a career. Junior developer. Content writer. Technical support. Data analyst. Customer service representative. Bookkeeper. These are not jobs that required no skill. They required exactly the kind of structured, rule-following, pattern-recognition work that large language models are now very good at.
The jobs at the top of the ladder are mostly intact, for now. Senior engineers are fine. Executives are fine. The people who spent a decade building expertise before AI arrived are largely fine. The people who were supposed to spend the next decade building that expertise are the ones getting cut before they can start.
What Getting It Right Used to Look Like
There was a specific version of success that Gen Z was sold. Go to university. Study something technical. Get a coding job or a writing job or an analytical role at a company with good benefits. Learn the industry. Move up. It was a slower path than the startup lottery, but it was supposed to be reliable.
The entry-level jobs that path depended on are contracting. Inside Higher Ed reported this week that roles requiring college degrees are among the highest-exposure categories for AI displacement. The credential that was supposed to be the moat is not a moat against a system that can write, code, and analyze at the same time, at scale, without needing health insurance.
Boston Consulting Group estimates that AI will reshape between 50% and 55% of U.S. jobs over the next three years. That word — reshape — is doing a lot of work in that sentence. It can mean restructured and improved. It can also mean eliminated and replaced. For entry-level workers, the direction it is moving right now is not ambiguous.
The Oracle Pattern Is Becoming Standard
Oracle sent 30,000 termination notices by 6am email while posting $6 billion in quarterly profit. The money was not gone. It was being redirected. Away from human labor. Toward AI infrastructure. That is the corporate math behind the 16,000-a-month number. Companies are not cutting because they are struggling. They are cutting because they are profitable enough to buy systems that replace the humans who generated the profit.
This is not a technology story. It is a distribution story. The gains from automation are going somewhere. The question is whether the workers who were displaced by that automation have any claim on where they go.
The Counterargument Exists and Is Insufficient
Bureau of Labor Statistics projections show software developer employment growing 15% by 2034. IBM says it is tripling entry-level hiring. Some economists argue that new jobs always emerge to replace the ones technology eliminates. They are not wrong historically.
But historically, that transition happened over decades. The agricultural-to-industrial shift took generations. The industrial-to-service shift took a generation. The AI transition is happening over years, in some sectors over months. The speed matters. A 25-year-old who cannot get their first job does not have the luxury of waiting for the labor market to rebalance over the next thirty years.
58% of workers who use AI in their jobs say their position is safe from AI entirely. That confidence is being built during the same period the models themselves are getting significantly more capable. What feels safe today has a track record in this environment of not staying safe for long.
What This Actually Means Going Forward
Gen Z is not the only generation that will feel this. They are the first. The exposure is moving up the skill ladder, not staying at the bottom. The entry-level positions are gone first because they are easiest to automate. The mid-level positions are being restructured now. The senior positions will be next, on a longer timeline, but the direction is not reversing.
The 16,000-a-month number is a current rate, not a peak. The models being deployed this year are more capable than the ones deployed last year. The organizations building them are not slowing down. The question is not whether the number continues. The question is what happens to the people inside it.