AI and Older Workers in 2026: What New Research Means for a Career Change at 45

AI is changing which older workers stay in their jobs, but the effect is uneven. A June 2026 Boston College study found that workers 55 and older in AI-exposed jobs now leave work more often than before ChatGPT arrived. The risk varies sharply by job. Your task mix matters more than your age.

If you are in your forties or fifties, this is a planning question, not a panic button. At Simply Forties, we cover the money and career decisions of midlife adults. This guide explains what the new research found, how to judge your own exposure, and how to move toward safer work without draining your savings.

What the Boston College Research Found About AI and Older Workers

The study comes from the Center for Retirement Research at Boston College, in a brief by Geoffrey Sanzenbacher. It compares job exits before and after ChatGPT launched in November 2022. The data come from the Current Population Survey, matched with an index that scores each occupation from 0 to 100 on AI exposure.

Before ChatGPT, older workers in AI-exposed jobs were less likely to leave work. Those jobs paid more and rarely required heavy lifting. After the launch, that advantage shrank, and exits into unemployment rose.

The author is careful to call the findings early. Hiring slowdowns, research funding cuts and the growth of the AI sector itself could all be pushing the numbers. Still, the direction is clear enough to plan around.

The table below shows two ends of the range the brief reported.

The brief also reports overall exit rates across the full period. They were 11.7% for highly exposed jobs and 14.1% for low-exposure jobs. In other words, workers in exposed jobs still leave less often on average. The change is in the trend, not the starting point.

Notice what the table does not say. Painters still leave work at a higher rate than programmers. Physical jobs wear bodies down, and that matters more than software for many workers in their late fifties.

There is a counterpoint worth stating. The same research notes that AI could also help older workers stay longer, by taking over the tiring or repetitive parts of a job. A 58-year-old analyst who uses AI to cut a day of spreadsheet work may have more energy for the parts that need experience. Which way it goes depends on whether the employer treats the tool as help or as a headcount replacement.

Is Your Job Actually Exposed to AI?

Your job is exposed if a large share of your weekly tasks involve drafting, summarizing, sorting data, routine analysis or answering standard questions. Those are the tasks current AI tools do well. Jobs built on physical work, in-person care, trust and judgment calls are exposed far less.

Job titles hide the real picture. Two people called "office manager" can have very different exposure. One spends the week on scheduling and invoice entry. The other spends it on vendor disputes, staff conflicts and safety checks.

Try a simple audit this weekend. List your tasks from last week and sort them into three groups.

If most of your week sits in the first group, you have a reason to start planning now. If most sits in the third, your risk is lower, but your employer's budget choices still matter. The Bureau of Labor Statistics (BLS) Occupational Outlook Handbook lists projected growth and typical pay for hundreds of jobs, which helps you compare your own role against the field.

Which Midlife Roles Face More Pressure?

Roles built mostly on screen-based processing face the most pressure. Think of bookkeeping entry, basic coding, claims processing, standard report writing and first-line customer chat. These jobs are not vanishing overnight, but employers are already asking whether one person with a tool can do the work of two.

Roles with a steadier outlook tend to mix human contact with specialized knowledge. Nurses, electricians, HVAC technicians, social workers and skilled supervisors fit that pattern. Their work needs a person in the room, a license, or trust built over time.

How Can a Worker in Their Forties or Fifties Respond?

The best response is to shift your task mix toward work that needs judgment, while learning the AI tools your field already uses. You do not need a full reinvention. Most people can move a few steps sideways within their industry and keep most of their pay.

Survey data cited in the Boston College brief show only 18% of workers aged 50 to 64 use generative AI. That gap is an opening. A worker who can use the tools and also bring twenty years of context is hard to replace.

Three moves tend to work well for midlife workers.

Talk to your manager early, too. Ask where the team plans to use AI over the next year and which skills the company wants to build. That conversation shows whether your employer sees you as part of the plan. It also gives you a head start if the answer is vague.

Our look at the fastest-growing jobs to 2035 shows where hiring is headed. Many of those roles sit in health care, skilled trades and technical support, where human contact still carries the work.

Paying for a Move Without Draining Your Savings

Retraining costs less than most people expect if you pick a short, targeted program. Community colleges often run certificates in months, not years, at a fraction of university tuition. The Department of Labor's CareerOneStop site lets you search local training providers and find your nearest American Job Center.

Money planning should come first. Our guide on how to pay for retraining at 45 walks through budgeting a gap year, using employer tuition help and avoiding high-interest debt. Test the new field cheaply before you commit, with a short course, a volunteer project or a part-time contract.

Leaving work early also changes your retirement math. Every year out of the workforce can lower your Social Security earnings record and your savings growth. Our explainer on the 2026 Social Security Trustees Report shows why those years carry weight when the program's trust fund faces a funding gap in 2034.

Bring your household into the plan as well. A partner, an adult child or a parent you support may all be affected by a pay cut during retraining. Agree on a floor for monthly spending and a date to review the plan. Short, concrete numbers beat vague worries every time.

Know Your Rights If a Layoff Comes

The Age Discrimination in Employment Act (ADEA) protects workers who are 40 or older. According to the Equal Employment Opportunity Commission (EEOC), it covers employers with 20 or more employees. You generally have 180 days to file a charge, though state laws may extend that window.

Read any severance offer slowly. Release agreements for workers 40 and older come with extra rules, including time to review the offer. Ask whether younger employees in the same role were kept, and keep copies of your reviews and emails. If you also lost your job search momentum, our piece on job searching after 45 covers how to shorten a long search.

A 90-Day Plan for the Next Step

A short timeline keeps worry from turning into delay. The plan below fits around a full-time job.

At the end of 90 days, you will know more than most workers your age know about both the tools and your own options. That knowledge is worth having even if you stay exactly where you are. Simply Forties editors recommend testing early, because a small trial makes the bigger decisions calmer.

Keep the plan small enough to finish. A modest plan that you complete beats a bold plan that stalls in week three.

Where This Leaves You

AI is not ending midlife careers, but it is reshuffling which ones last. Audit your tasks, learn one tool, and test a new direction on a small scale. Keep your savings and your Social Security record in view while you decide. Simply Forties will keep tracking the research as new data arrives.

— 40 —

Some links may be sponsored. Our editors choose what we cover, and nothing here is financial, legal or benefits advice. Our editorial policy.