MIT economist Frank Nagle has a blunt warning for executives racing to trim payrolls in the name of artificial intelligence: they are sabotaging their own future.

Nagle, who also serves as advising chief economist for the Linux Foundation, studies how AI is reshaping day-to-day work. His research tracks who does what inside companies, how long tasks take and how those patterns shift once AI tools are introduced.

He argues that jobs are sorting into three broad groups. At one end are roles where AI can handle core tasks almost entirely on its own. Translation is his clearest example, powered by vast troves of multilingual text. Yet even here, humans remain essential as fact-checkers and specialists in underrepresented or spoken-only languages.

At the other end are physical, hands-on jobs that current AI systems cannot easily touch. The workers building a house outside his office window, he notes, are doing tasks that software alone cannot automate. Robotics may eventually change that, but not at scale today.

The largest and most important category sits in the middle: jobs that survive but change profoundly. Software developers are his prime case study. When AI coding tools arrive, developers do more coding and less project management. The effect is especially strong for junior staff, whose coding time jumps severalfold compared with senior colleagues.

That is why Nagle calls it a “critical strategic mistake” to cut entry-level roles and claim AI as the justification. Junior employees, he says, both gain the most from AI and generate the most insight into how workflows should evolve. They are also the pipeline for tomorrow’s leaders. Firms that keep hiring will have access to “more and possibly better people” as rivals pull back.

Nagle believes many headline-grabbing AI layoffs mask more mundane problems: overhiring, overexpansion and poor management. AI, he says, is an “easy scapegoat” for decisions that are really about budgets, not technology.

Looking ahead, he urges students to pair technical skills in AI or computer science with humanities training that sharpens judgment about why and how technology is used, and to consider healthcare, where demand is rising even as AI spreads.

For educators and employers alike, his message is consistent: assume people will use AI, teach them the fundamentals anyway and train them to spot when the machine is wrong. The winners in the AI era, Nagle suggests, will be those who redesign work around humans plus machines, not humans versus machines.