Will AI Take My Job? What Boston College Found About Gen X and Workers Over 55

If you've got a knot in your stomach about where work is headed, you're not imagining it.

Boston College just put some numbers behind that feeling. Workers 55 and older in the jobs most exposed to AI are leaving work more often than they used to, and a lot of them are landing in unemployment, still looking. Meanwhile, young people are having a hard time getting hired in the first place.

We're getting squeezed from both ends.

I know a little about getting squeezed out.

I was a mechanical draftsman at the same company for 17 years. By the end I'd mentally checked out, quiet quitting before anybody called it that, and on July 3rd, 2013, they let me go. What kept me afloat was the stuff I'd been tinkering with on the side.

So when a study like this comes out, I read it slowly.

The study comes from the Center for Retirement Research at Boston College.

Geoffrey Sanzenbacher used Current Population Survey data to track workers 55 and older one year apart, from 2014 to 2025, skipping the COVID years. He matched their jobs to an AI exposure index from Tufts and compared job exits before and after ChatGPT launched in November 2022, with controls for things like education and earnings. In the model's predictions, painters see "just a 2-percent increase" in leaving work. For computer programmers, "the increase is over 25 percent." Jobs in the middle land somewhere around "a 9 percent to 16 percent increase in exit."

Those are not small numbers.

Now the fine print, because it matters.

The rise is "specifically to unemployment (but not out of the labor force)." In plain English, these folks lost a job and are still looking. Nobody's being forced into retirement by a chatbot. These are model predictions from an early look, they show a pattern rather than proof, and the author himself says "caution is in order." White-collar jobs still have lower exit rates than physical ones, but that old advantage has "been greatly reduced."

The cushion we thought our desk jobs gave us just got a lot thinner.

Stanford is seeing the other end of the squeeze.

Erik Brynjolfsson and his team looked at ADP payroll data covering millions of workers. Employment for 22 to 25 year olds in AI-exposed jobs "now stands 19% below where it would be" if it had kept pace with their less-exposed peers, mostly through "reduced hiring of young workers." Meanwhile, "experienced workers show no comparable gap." They found "no evidence of widespread, economy-wide job displacement," and where AI helps people do their work, employment is "flat or rising, especially for experienced workers." The authors are clear these are early, descriptive signals and not proof of cause.

Young people can't get in the door, some of us in the most exposed jobs are getting pushed out of it, and experienced folks using these tools as a helper are mostly holding steady.

The tools are good, but they're not always right.

Harvard and BCG ran an experiment with 758 consultants. On tasks AI handles well, the ones using it finished 12.2% more tasks and worked 25.1% faster. On a task picked to be beyond what AI could do, they were "19% less likely to produce correct solutions." The group without AI got it right about 84.5% of the time, while the AI groups scored 60% and 70.6%.

Knowing when it's wrong is the skill, and that's where decades of doing the work come in.

Most of us are just getting started with this stuff.

AARP found AI use among adults 50-plus has nearly doubled, "rising from 18% in 2024 to 30% in 2025." In another AARP survey, only 7% of adults 50-plus called themselves advanced users and 1% said expert. That's a lot of experienced people who haven't paired their judgment with these tools yet.

Which means there's still room to get ahead of it.

You don't need a five-year plan for this.

You need one month and three small steps. Each one builds on the last, and none of them require quitting anything. Here's what I'd do if I were starting this month.

Week 1: Audit what you actually do.

List the tasks you do in a normal week and mark each one: AI can do it, needs my judgment, or mixed. Then run the Upwork worth check I wrote about a while back. Write down your past jobs, projects, software, hard skills and soft skills, then search Upwork's "Talent" section to see what people charge and its "Jobs" section for 3 to 5 of your skills to gauge demand. I made a short video walking through it.

It's a quick reality check on what your experience is worth outside your current job.

Weeks 2 and 3: Pair up with one tool.

Pick one AI tool and use it on one of your mixed tasks every day. Keep a simple log of where it got something wrong and how you caught it. That log is proof of your judgment, which is exactly what the Harvard study says matters. As I put it in an earlier post, "You need your experience in a form a helper can use."

Two weeks of that and you'll know where you beat the machine.

Week 4: Own one skill in public.

Pick one niche skill and make one short video about it. Record it in Riverside, let AI handle the rough edit, and publish in 30 to 60 minutes. Then post it across a few platforms (I schedule TikTok and Instagram straight from Riverside). People who need that skill start finding you, and you send them to an email list you own, because social is rented land.

That's the whole system.

It works for people who show up consistently. Kat Norton, aka Miss Excel, posted her first TikTok in June 2020 and "began posting once a day," teaching Excel. According to CNBC, she grew to "more than 1 million followers across TikTok and Instagram." She was 28, so I'm pointing at her method here.

I'm testing this daily-video system myself right now on a side project, and I'll tell you how it goes.

60 days, three steps, and you come out the other side with proof.

Have a good one,

Corey