AI taking jobs has been the most talked-about topic in American business for the past two years. Announcements of layoffs were made grimly frequently, with automation being blamed for at least some of them. Employees became nervous after reading the headlines. The majority of Americans, at 53%, are now concerned that AI may result in a loss of employment for them or a family member. It’s a legitimate fear based on actual occurrences. Additionally, new research indicates that it’s not quite the whole picture.
Between 2021 and the beginning of 2026, a study conducted by Ramp Economics Lab and Revelio Labs tracked employment and AI spending in about 22,000 U.S. businesses. The narrative is contradicted by the headline finding. Over the course of two years, companies that spent the most on AI—an average of roughly $34 per employee per month—grew their total headcount by 10.2%. Low-intensity adopters saw virtually no change at all, spending closer to $3 per employee.
What happened to entry-level hiring is more difficult to ignore. The idea that AI would erode the bottom rung of the career ladder and that recent graduates and young workers would find their doors closed before they even knocked is where the fear has focused. Over the same time period, entry-level headcount increased by 12% for high-intensity AI adopters. Teams in engineering increased by more than 7%. Customer service, sales, and finance are all included. It’s the type of information that doesn’t neatly fit the narrative and merits more than a brief assurance.
But there’s a catch, and it’s important. The businesses making the biggest investments in AI weren’t chosen at random. Before they even conducted their first AI pilot, they were already bigger, more technically advanced, and expanding more quickly. Because of this, causality is actually ambiguous. Did AI fuel the hiring boom, or were these businesses just early adopters with greater resources and momentum? In all honesty, the researchers admit that their comparison of early adopters and later adopters in comparable industries is cleaner than a direct comparison, but it is still inconclusive. That question cannot be answered by a single study.
The data does point to a sort of lag effect. Hiring hardly changes during the month that a company makes significant use of AI. It begins to change six months later. The difference between high adopters and everyone else is evident by the time they are twelve to eighteen months old. The study’s lead economist put it simply: most businesses utilizing AI today aren’t seeing results yet, and that’s likely because they haven’t crossed the line that separates a corporate experiment from a real deployment. It appears that increases in productivity come before increases in headcount, not the other way around.
There is a certain logic to that framing. Even if a business uses AI agents to find potential customers, someone is still needed to answer the phone and complete the transaction. In order to take advantage of the new opportunities that speed opens up, a software team that ships more quickly still needs more engineers. In these situations, AI appears more like a force multiplier than a replacement, and multipliers typically need more people to absorb what they produce. As this develops, it’s difficult to ignore the fact that the businesses that are still operating in pilot mode are neither making money nor losing it. It appears that half-measures yield half-outcomes.
Nevertheless, it would be too simple to conclude from this study that the AI employment crisis is resolved. Salesforce, Amazon, and Meta have all directly linked layoffs to the use of AI. White-collar, tech-forward companies make up a large portion of the research sample. It doesn’t account for the experiences of truck drivers, warehouse employees, or anyone whose work involves anything other than engineering and enterprise software. According to estimates from the World Economic Forum, by 2030, new technology will create 170 million jobs while displacing 92 million others. On paper, this is a net positive, but for workers in the displaced column, the numbers are cold comfort.
The more accurate picture is that AI is changing the nature of work before anyone fully understands how, which is what significant technological changes typically do. Jobs are evolving more quickly than data can keep up. Compared to three years ago, entry-level positions at high-adoption companies are now more technical, flexible, and reliant on the ability to work with tools that weren’t around yet. These days, a clerical worker requires greater technical literacy. Stronger interpersonal skills are essential for software developers. Even though the floor is shifting rather than vanishing, standing on it can still be challenging.
The businesses investing the most in AI are hiring. That is true. However, it’s not quite as reassuring as it first seems, and anyone who presents it as a straightforward happy ending is probably trying to sell you something else as well.
