Every week brings new headlines about AI destroying jobs. But labor economists who study the actual data are finding something more complicated, and in some ways more alarming. The threat is real, but it’s landing in very specific places, on very specific people. Here’s what the numbers actually say.
Key Takeaways
- Unemployment in AI-exposed jobs is currently lower than in less-exposed occupations. But entry-level positions dropped 16% in one year.
- Only 1 in 5 companies currently uses AI in any business function
- The disruption is just beginning: wages in AI-exposed sectors are rising, but opportunities for young workers are shrinking fast
What the Data Actually Shows
Here is a fact that surprises most people: workers in jobs most exposed to AI currently have lower unemployment rates than workers in jobs least exposed to AI. That sounds backward. If AI is replacing workers, shouldn’t those jobs have higher unemployment? The answer reveals why the current moment is so difficult to read, and so easy to misread in both directions.
Labor economist Erika McEntarfer has put it plainly: “All of the available evidence to date suggests that AI’s impact on current labor market conditions is likely small right now.” The disruption that fills headlines every week is real and coming. But right now, in 2026, most of it is still forecast, not experienced. The jobs haven’t disappeared yet. The conditions are changing.
Only 1 in 5 companies currently uses AI in any business function. If AI were already reshaping the labor market broadly, you’d expect that number to be much higher. Most businesses are still figuring out how to use these tools, which means the wave hasn’t broken yet. The disruption described in headlines is largely a preview, not the main event.
Wages in AI-exposed sectors have actually risen relatively fast since ChatGPT launched in late 2022. Employers are still paying up for experienced workers in these fields, because AI tools still require skilled humans to direct them, verify their outputs, and build on their capabilities. For experienced professionals, the AI wave hasn’t been a threat. It’s been a productivity amplifier.
Where the Real Pain Is Already Happening
The picture changes completely when you look at where people enter the workforce. Stanford researchers tracked entry-level job postings in AI-exposed fields from 2024 to 2025. They found a 16% decline in one year. That number is not small. It means that for every 100 junior roles that existed before, only 84 exist now. The first rung on the career ladder is being pulled up.
Recent college graduates are feeling this directly. Unemployment among new graduates currently sits at around 5.6%, the highest it’s been since the pandemic, and before that, since the 2008 financial crisis. The jobs that graduates used to take to build their skills and accumulate experience are the exact jobs most likely to be partially automated first: data entry, basic coding tasks, customer analysis, content drafting.
The coding job market, which many people assumed would be AI-proof because of the skills required, tells a similar story. Overall coding employment is still growing. But the growth rate has slowed by about 3% since ChatGPT launched. That 3% gap, spread across a huge market, represents tens of thousands of positions that would have existed in a world without AI tools.
The pattern that emerges is consistent: AI is not yet eliminating careers in bulk. It is quietly reducing the number of people needed to do the foundational work in a field. Senior professionals remain valuable. Junior roles (the training ground for every career) are quietly disappearing.
Also on Save Your Job:
- Devin Writes 89% of Its Own Company’s Code. You’re Next.
- AI Backlash Is Here: Graduates Boo AI Speakers on Stage
- Entry-Level Jobs Crisis: AI Is Closing the Door on Graduates
What This Means for Your Career Right Now
The “AI’s impact is small right now” framing from economists is technically accurate but practically misleading for anyone building a career in 2026. Here’s why: the entry-level positions that no longer exist were not just jobs. They were the apprenticeship period that turned new graduates into experienced professionals. When those positions shrink, the pipeline for developing the next generation of skilled workers shrinks with them.
The companies that are currently paying well for AI-exposed roles are paying for seniority and judgment, things that come from doing foundational work for years. If the foundational work is automated before the next generation gets to do it, where does that experience come from? Nobody has a clean answer to that yet. But waiting for the economy to figure it out for you is not a strategy.
Over the next three to six months, watch for these patterns to accelerate, not reverse. The 1 in 5 companies using AI today will become 2 in 5, then 3 in 5. Each adoption wave will compress entry-level roles further. The workers who adapt proactively (by learning to work with AI tools rather than alongside them) will be the ones whose roles survive and grow.
The data is clear: the disruption is targeted, it is accelerating, and it is concentrated at the beginning of careers. Experienced workers who understand what AI can and cannot do are thriving. Workers who wait for clarity before adapting are the ones who will look back and wish they had started earlier. If you want to understand what AI means for your specific role and how to stay ahead of it, start at Save Your Job From AI.
Follow the story on Save Your Job.
About the Author
