Can We Really Predict Which Jobs Are Safe from AI?

According to the BLS, nurse practitioners will see the most growth in the job market in the next ten years. The demand for this role is projected to grow by 41% between 2025 and 2035. 1

Other healthcare jobs appeared prominently on the BLS’s list, including medical and health services managers and physical therapy assistants. Relatedly, healthcare is now the largest employment sector in the US, with over 18 million Americans employed in the industry. 2

Many experts have posited that the jobs that will dominate the US economy in the future will be AI-heavy, given the push by many companies to implement the technology into their workflows.

Healthcare’s prominence in the BLS’s list cuts against this expectation somewhat; many jobs in this industry are very human-centric and have little to do with AI.

Why are healthcare workers in so much demand?

In the majority of Western countries, including the US, the population is aging rapidly.

According to Georank—drawing from data from the World Bank, United Nations, and the US Census Bureau—the median age in the US is predicted to reach 41.4 by 2035. 3 In 2018, the US Census Bureau predicted that 65-year-olds would outnumber 18-year-olds by 2035. 4

Naturally, older people require healthcare services more than younger Americans. Assuming these demographic projections are correct, the healthcare industry will need an influx of new workers to meet the demand.

A lack of willing workers

This presents a potential problem for the labor market, and, more generally, for the country itself. Healthcare jobs pay well, but it’s been tricky to get American workers on board, especially members of Gen Z.

According to recent findings from the Harris Poll, 70% of Gen Z healthcare workers are planning to quit their jobs within the next year. Those polled claimed that a lack of career progression was responsible, with only one quarter reporting that they trust that their employer has their long-term career growth in mind. 5

For the healthcare industry, this lack of willing workers is troubling. The National Center for Health Workforce Analysis predicts a scarcity of about 500,000 healthcare workers by 2038. 6

The current data suggests that if you’re someone who’s looking to pick a career that will weather the AI-related shocks that the next decade will likely see, the medical field is a safe bet. If anything, right now, most experts actually expect nurses and other healthcare workers to have too much work.

But these projections don’t guarantee that workers in these jobs will be safe. So far, predicting which jobs will be insulated from AI-induced displacement has been tricky.

Side Hustles spoke with Sean Campbell, CEO of the B2B market research and AI advisory firm Cascade Insights. Campbell told us that most predictions “focus on current AI capabilities rather than its trajectory.”

“The minute you stop and try to make a meaningful prediction, it’s very easy for it to be off within a few weeks or a few months, with AI’s rate of advancement,” he said.

AI’s rapid rise

This is one of the central issues when it comes to evaluating AI’s impact on the labor market. Historically, technological advances have destroyed certain jobs, but also created new opportunities, and workers have mostly adapted accordingly.

However, Campbell argues that the rate of AI’s advancement makes it hard for workers to switch gears fast enough.

The problem, he said, is that in the past, “technological transformations would tend to take decades in most cases, giving people time to switch into those new occupations.” Conversely, AI’s rise is “happening quickly enough that people simply can’t re-skill fast enough to move into those new roles.”

“I think we’ll see a gap between those who can make the switch quickly and effectively and those who can’t,” he added.

Coding: the job once seen as the safest bet

Campbell’s view that it’s hard to know which jobs will surge or decline in demand has some evidence behind it. It’s particularly apparent in the field of software development.

During the 2010s, coding was believed to be one of the few jobs that would be safe from automation. Most of the job-loss discourse centered around blue-collar roles, with experts claiming that these workers would be displaced by robotics. When those in manual fields complained, the retort from many white-collar workers was that they should “learn to code.”

Fast forward to 2026, and the tables have turned. According to Anthropic’s Labor Market Impacts of AI report, which looked at the jobs most exposed to AI, computer programmers had the most observed exposure of any occupation (74.5%).

Conversely, blue-collar work is seeing a resurgence. A joint study from Deloitte and The Manufacturing Institute found that the manufacturing sector will need 3.8 million new employees between 2024 and 2033. 7

Anthropic’s report also pointed out that people with physical, hands-on roles, such as cooks and motorcycle mechanics, have the least exposure to AI. 8

Along with the jobs that are growing fastest, the BLS also listed the occupations that will see the most dramatic declines: 9

Most of these projections are no real surprise. The markets for switchboard operators and telemarketers, for instance, have been declining since before AI was on the scene.

However, other job losses are almost certainly due to AI. At the top of the list are word processors and typists, with a projected decrease of almost 35%.

These jobs involve converting rough notes, audio, or handwritten drafts into professional documents. Many typists work in transcription, translation, or captioning for movies and TV shows.

For years, typing jobs were a go-to remote side hustle for people without extensive experience or qualifications. Now, AI is becoming increasingly capable of handling writing tasks, and businesses are adapting accordingly.

A joint study from Harvard and Imperial College studied job listings on an undisclosed “leading freelance platform.” The researchers found there was a 21% drop in job posts for automation-heavy work, including writing, within eight months of ChatGPT’s launch in 2022. 10

The AI-induced disruption in the job market for coders and typists might seem inevitable in retrospect, but as Campbell pointed out, it wasn’t obvious at the time.

This, in turn, casts doubt on just how good an idea it is to pivot to a career in an in-demand field, like healthcare. Is there really any guarantee that you’ll be safe in the long run? Just how immune to AI and automation are the seemingly “safe” jobs?

Healthcare jobs

This industry isn’t immune to encroachment from AI, particularly where it overlaps with other more vulnerable classes of work.

For instance, Campbell told us that he believes those working in medical transcription are in for a rough ride as AI takes over many of the routine admin tasks they have traditionally been responsible for.

“If you’re a medical transcriptionist right now, I would be looking at a different career,” he said. “No one is going to pay more for their medical appointment so it can be human-transcribed.”

Nurse practitioners are unlikely to be ousted by AI, but there is research to suggest that nursing robots may eventually be able to take over some of the tasks human practitioners are responsible for. A study from the School of Nursing at Duke University in North Carolina proposed that nursing robots could handle physical tasks, such as: 11

• Transporting equipment
• Moving heavy equipment
• Setting up and maintaining sterile environments

While the study acknowledged the need for human practitioners, the authors stressed the need for nursing educators to incorporate robotics and AI into their training curriculums.

Manual jobs

Again, jobs in manual trades like construction are seen as some of the safest occupations from AI displacement. Many young people have even started pivoting to blue-collar work as technology automates office roles.

However, it’s unlikely that workers in these fields will be as secure as nurse practitioners. Campbell believes that young people “absolutely will be at risk” from automation, and sees the current surge in demand for blue-collar work as temporary.

“I think, long term, blue-collar jobs are perhaps more at risk if we include robotics in the equation,” he said. “In the near term, say several years to a maximum of 10 years, they’re probably safer jobs. If you’re a young person looking for a 50-year career in the trades, I don’t know if I would make that bet.”

The humanoid robotics market is projected to reach $38 billion by 2035, according to Goldman Sachs. Their report states that manufacturing is particularly vulnerable to automation, as evidenced by the application of robotics in China. 70% of the work in the manufacturing field in China is already done by machinery, and only 30% is handled by humans (and a third of this consists of managerial work, not physical labor).

A paper published in the University of Michigan’s Journal of Economics argued that blue-collar workers aren’t at an immediate risk of displacement, but stated that in the long term, as AI makes its way into the physical world, people in hands-on jobs will likely be affected. 12

Whether you’re a blue-collar or white-collar worker, the future is hazy. Even those in fields seemingly insulated from AI could see technology change their jobs or replace them altogether.

When the future is this challenging to predict, how can workers across industries protect themselves? For the future workforce—those still in college—what’s the best way to ensure you’re studying something that will remain useful in a decade?

“I think you need to look closely at the skills you’re learning and how durable they are,” Campbell told us. “Some programs teach skills that go beyond just rote memorization or learning a set of topics that AI might eventually conquer.”

In practice, this, he suggested, means honing soft skills like adaptability and critical thinking rather than over-focusing on any one set of hard skills.

You should be “learning how to develop as a person to increase your ability to judge things appropriately and develop your intuition and your creative skills,” he said. He and many other experts believe that those innately “human” qualities are beyond AI’s capabilities.

Of course, as we’ve seen time and time again, what people believe AI is incapable of often turns out to be what it excels at. For this reason, Campbell also suggested that young workers make it a priority to broaden their skills as much as possible. Put another way, diversify instead of specializing.

“Will it matter to develop deep competence, or will it be better to be more of a polymath?” he said. “I would argue that polymaths will tend to win out in the future.”