🛟 Are there any "Safe Careers" Left?

WorkLifeFuture · Guidance for parents supporting emerging adults in the age of AI & automation

🎇 Happy Tuesday,

When I began my journey into corporate America decades ago, the career advice was pretty straightforward:

Get a solid degree. Get your foot in the door at a reputable company. Work hard. Move up.

And it was pretty sound guidance.

It isn't enough anymore.

I don't say that to alarm you. I say it because a new report landed on my desk this week.

The Center for an Urban Future just published a study of entry-level jobs in New York City using data from labor market analytics firm Lightcast and AI exposure measures from the Anthropic Economic Index.

Here's what they found:

That last number should get our attention. One of the fields parents and colleges have been pointing kids toward for decades just saw its entry-level pipeline in New York nearly cut in half in three years.

Now, this is a New York study, so we shouldn't pretend it represents the entire country. But it is a pretty loud warning from one of America's biggest white-collar labor markets.

The jobs themselves haven't necessarily disappeared. The bigger problem may be the bottom rung.

And that is the rung our kids need to get started moving up.

AI exposure does not mean a job is doomed. It means a meaningful share of the work can increasingly be assisted, accelerated, or automated by AI.

Think:

When more of a job is made up of those tasks, companies may need fewer people to produce the same amount of work. That doesn't necessarily mean mass unemployment. It can mean something subtler:

Fewer entry-level hires, higher expectations for each worker, and a much harder first step onto the career ladder.

That distinction matters. At the same time, work that depends on physical presence, hands-on skill, direct care, judgment, trust, or unpredictable real-world environments is harder to automate.

AI may still help. But it doesn't replace the core of the job.

The Bureau of Labor Statistics' new projections through 2035 reinforce an important point: there is no simple "white collar bad, blue collar good" story here.

Some of the fastest-growing occupations are in healthcare and energy infrastructure. Nurse practitioners are projected to grow roughly 41%. Solar installers about 37%. Electricians about 9%, with tens of thousands of openings each year.

But data scientists are also among the fastest-growing occupations, and software developers still have strong overall projected growth. Meanwhile, computer programmers, as a standalone category, are projected to decline.

That's why broad labels like "tech" aren't very useful anymore. You have to look underneath the title.

What does the person actually do? Which parts of that work are becoming easier to automate? And maybe most important, can a young person still get hired at the bottom?

If you were sitting down with an 18-year-old today, how would you evaluate a career?

I'd start here.

Don't ask whether AI exists in the field. It exists everywhere.

Ask how much of the actual daily work consists of tasks AI can increasingly do faster or cheaper. A software developer may still have an excellent career, but writing boilerplate code, basic debugging, documentation, and routine testing are exactly the kinds of tasks AI tools are getting better at.

That changes what employers need from junior workers.

Some demand comes from trends that are hard to reverse. An aging population increases the need for healthcare. Electrification, data centers, EVs, housing, and grid investment require skilled workers.

Demand driven by demographics, infrastructure, and physical needs can be more durable than demand created by a temporary business cycle.

A nurse physically cares for patients. An electrician works inside buildings. An HVAC technician diagnoses a broken system in a real house on a 105-degree afternoon.

Those jobs cannot simply be moved into software. That is an advantage.

A trial lawyer needs to read a room. A nurse needs to calm a frightened patient. A good salesperson understands what a customer actually means. A manager has to make decisions with incomplete information.

AI may support these jobs, but the human part remains important.

I call this kind of work "anthrosive." These are jobs that are better handled by a human. For now.

And young workers who can do this work using AI are at a huge advantage, too.

This may be the question families overlook most. A career can have strong long-term demand and still be a difficult place for a 22-year-old to start.

That's what makes the current tech numbers so important. Experienced workers may remain highly valuable. But if AI allows them to do more, companies may need fewer junior people underneath them.

A great career is a much tougher bet if the apprenticeship ladder has broken.

A licensed electrician might work for a contractor, specialize in solar or EV infrastructure, move into supervision, become an inspector, or start a business. A nurse can move among specialties, settings, management, education, or advanced practice.

A strong skill should create more than one possible future.

Optionality matters. Like my buddy, Paul: a software engineer and part-time licensed electrician.

These aren't prescriptions. Every kid is different.

But they show why we need to think differently.

Software still has strong long-term demand. But entry-level tech hiring has fallen sharply at the same time AI coding tools are allowing experienced engineers to do more.

That doesn't mean software development is a bad career. It means the first step may be getting harder.

And I can tell you from personal experience: I work with a group of brilliant developers who have maximized their capabilities with traditional coding and reasoning skills turbocharged by AI tools.

Electricians face a very different problem. The work requires physical presence, licensing, apprenticeship, and hands-on skill. Demand is being supported by housing, grid modernization, EVs, solar, data centers, and infrastructure.

One field may have strong demand at the top but a narrowing on-ramp. The other still has a formal apprenticeship ladder.

That difference matters.

A junior financial analyst may spend a large part of the day building models, gathering data, summarizing information, and producing reports. Those are exactly the kinds of tasks AI is improving at rapidly.

A registered nurse operates in a very different environment. The work requires physical presence, judgment, care, trust, communication, and real-time response to unpredictable situations.

The BLS projects more than 180,000 nursing openings per year on average.

The point isn't that everyone should become a nurse. It's that the structure of the work matters.

I am not telling you to pull your kid out of computer science. And I am not telling you to push them into the trades.

The point is simpler.

The old shortcut ("good school + practical major + good company = safe career") doesn't work as reliably as it once did.

We need better questions. What does the work actually look like? Which parts can AI do? Is demand growing for lasting reasons? Can my kid realistically get hired at the entry level? Does the career lead somewhere? And does the underlying skill create options?

That is a much better conversation than simply asking, "What majors pay well?"

Try this conversation with your emerging adult this week:

"I read something interesting about how AI is changing which entry-level jobs are growing and shrinking. Not the scary version. Just the data. Want to look at it together?"

Then pull up the Center for an Urban Future report or the BLS employment projections.

Let them react. Let them disagree. Let them ask questions.

Don't steer.

The goal isn't to hand them an answer. It's to help them develop the habit of evaluating careers based on where the work is going, not where it has been.

If this changed how you think about career advice, please forward it to one other parent who might still be operating on the old playbook.

👉 worklifefuture.com/subscribe

Don't help your kid choose a career because it looked safe ten years ago.

Help them think of one based on where the work might be ten years from now.

That's part of the new job of parenting through this.

Barry

WorkLifeFuture helps parents of emerging adults navigate the AI era together. Barry Kruse is the Head of Customer Success at Symbolic.ai, an AI writing and editorial platform for newsrooms and corporate communications teams. WorkLifeFuture is a personal project and does not represent Symbolic.ai's views.