For the past year, the story about AI and jobs has swung between two extremes. First it was mass job loss. Then, as hiring numbers ticked up and CEOs softened their language, the story flipped to AI being good for jobs after all. A new monthly tracker from workforce research firm Revelio Labs is the first serious attempt to check which version is actually true, and the answer is neither.
The tracker looked at companies since ChatGPT launched in late 2022 and split them into two groups: those that have genuinely adopted AI tools and those that have not. The AI adopters grew their headcount 27 percent faster than non-adopters. That number alone supports the optimistic story, and it is the number most people will quote.
The problem is who that growth actually reaches. Senior roles at AI-adopting companies grew 31 percent. Junior roles grew only 6 percent. At the same time, hiring demand for jobs that AI can do a lot of, like data engineers and financial analysts, dropped 36 percent compared to jobs AI barely touches. So the companies winning with AI are hiring, but mostly hiring people who already have experience.
This lines up with what is happening outside the tracker too. Recent college graduate unemployment has climbed to around 5.6 percent, well above the rate for college graduates overall. Nearly a quarter of recent grads are working jobs that do not require a degree at all. Separately, layoffs tied to AI have already passed 100,000 for this year alone, through the middle of the year, which is nearly double last year's full-year total. Hiring headlines and layoff headlines are both true at once, just for different groups of workers.
There is a second finding in the tracker worth paying attention to, maybe more than the headcount numbers. The pace at which day-to-day tasks inside jobs are changing has doubled over the past several months. Almost all of that change is happening inside existing job titles, not by people moving between jobs. A marketing coordinator is still called a marketing coordinator, but what fills their day has shifted fast. Job titles are a lagging indicator. If you only track headcount by title, you will miss most of what AI is actually doing to work.
For a business owner, the practical read is this: do not assume a hiring rebound means entry-level hiring is back too, because right now it is not, at least not broadly. If you run a company that has actually built AI into how work gets done, you are likely growing, but the growth shows up as senior hires and reshaped roles, not as a wave of new junior positions. And if you manage a team, the bigger risk to watch is not that a role disappears. It is that the role stays on the org chart while the actual work inside it changes shape without anyone updating the job description to match.