What the Research Says About AI and Your Career After 40

Checked against primary sources by the Editorial Team. Every figure in this article was traced to the study it came from, including the ones that contradict what we previously published. How we review.

This article contains affiliate links. If you buy something through one of them, we may earn a commission at no extra cost to you. It never changes what we recommend or what we say about it. Advertiser Disclosure.

If you are in your 40s or 50s, the headlines about artificial intelligence can leave you anxious and tired. Every week brings a new prediction about which roles will vanish. This article sticks to what four primary sources report, and it flags where those sources stop.

What the biggest jobs report projects

Our earlier version of this article quoted the World Economic Forum as projecting that 70% of the skills required for most jobs will change by 2030. We could not find that figure in the Future of Jobs Report 2025. The report, published January 7, 2025, says: "On average, workers can expect that two fifths (39%) of their existing skill sets will be transformed or become outdated over the 2025 to 2030 period."

The same report projects "the creation of new jobs equivalent to 14% of today's total employment, amounting to 170 million jobs," and the "displacement of the equivalent of 8% (or 92 million) of current jobs." That works out to "net growth of 7% of total employment, or 78 million jobs." These are employer survey projections. They describe the whole labor market, and they do not tell you what will happen in your role.

Tasks change before jobs do

Researchers at OpenAI and the University of Pennsylvania studied how large language models line up with U.S. occupations. The authors report that about 80% of the U.S. workforce could have at least 10% of their work tasks affected, and that around 19% of workers may see half or more of their tasks affected. They also report that the effects span all wage levels, with higher income positions potentially more exposed.

That last point matters if you hold a senior role. Seniority does not exempt anyone from exposure. The authors measured how much work AI could touch. They did not project timelines or job losses, and they say so.

What the payroll data shows about age and experience

A study by Erik Brynjolfsson, Bharat Chandar, and Ruyu Chen at the Stanford Digital Economy Lab used payroll records from ADP covering millions of U.S. workers through June 2026. The authors report no evidence of economy wide job displacement. They also report that "employment of young workers (ages 22 to 25) in AI exposed occupations now stands 19% below where it would be had it kept pace with that of their less exposed peers." On experienced workers, they write that these workers "show no comparable gap."

Read that carefully. The study did not compare people over 40 with people under 40, and it did not test why experienced workers held steady. It does show that, so far, the hit has landed at the entry level. Anyone who tells you the research proves a midlife advantage is going further than the data goes.

What happens when people use the tools at work

Brynjolfsson, Danielle Li, and Lindsey Raymond studied a generative AI assistant rolled out to customer support agents. The working paper reports that the tool "increases productivity, as measured by issues resolved per hour, by 14% on average," with "a 34% improvement for novice and low skilled workers but with minimal impact on experienced and highly skilled workers." The peer reviewed version appeared in The Quarterly Journal of Economics in 2025.

Two lessons follow, and both come with limits. First, the tool helped most where workers had the least experience, which is consistent with experienced people already holding much of what the tool provides. Second, this was one company, one job, and one tool. The authors do not claim it generalizes to a manager, a nurse, or an engineer.

One way to act on these findings

Based on the payroll and productivity studies, one reasonable way to act is to get hands on experience with AI tools in your own work, so you judge what they do well and badly from use, not from headlines. The research above describes exposure and productivity effects. It does not test a particular book, course, or product, and none of these sources says that any item will protect a career.

If you prefer to learn from a structured guide, one option is a book by a Wharton professor who has studied AI in the workplace.

What we removed from the earlier version

The earlier article stated that the skills AI cannot replicate "tend to develop with age and experience," that curiosity is "one of the strongest predictors of successful career transitions at any age," and that spreadsheets in the 1980s favored accountants who understood accounting. We found no source for these in the references we checked. The 2025 review of career adaptability literature we had cited (Sharma and Kaur, 40 studies from 2013 to 2024) lists predictors such as personality traits, self efficacy, hope, and social support. It defines curiosity as one dimension of career adaptability. It does not report curiosity as a leading predictor, and it does not analyze age differences. We have taken those claims out.

Judgment, stakeholder management, and mentoring may well matter more as routine tasks get automated. That is a reasonable hypothesis, and we did not find a study here that tests it.