AI-proof jobs that pay $100K: How to screen durability

The premise needs a correction before it can be useful

No job is literally immune to automation. That matters here because the phrase “AI-proof jobs that pay $100K” promises a cleaner future than the labor market is offering. Brookings published research in late 2025 using patent data back to 1850 and occupational task modeling, and the pattern is awkward for anyone hoping higher pay alone buys safety.

Over the last century, technological change has tended to hit lower-education, lower-pay jobs harder. Applied to AI, the same model suggests the occupations most exposed may have the opposite profile, with higher compensation and more credentials. That does not mean every well-paid job is suddenly at risk. It does mean the old folk wisdom, earn more and worry less, needs a rewrite.

The more useful question is which six-figure careers are structurally durable. Not immune, just harder to dislodge. The U.S. median annual wage across all occupations was $49,500 in May 2024 (BLS Occupation Finder, August 2025). Jobs that clear $100,000 are already in rarefied air, but pay alone is only part of the story. The other part is whether demand is growing, or whether retirement and turnover keep openings flowing even when employment is flat.

How to judge whether a six-figure job is actually durable

The first test is growth. The BLS projects employment over a ten-year window, and the average growth rate for all occupations is 3 percent (BLS Occupation Finder, August 2025). A job growing at 20 percent or 30 percent is playing a different game entirely. But growth can flatter. A field may expand quickly and still offer relatively few openings if the workforce is young and stable.

That is why annual openings matter just as much. BLS separates projected openings from percent change, and those openings include new jobs plus positions vacated by workers leaving the labor force. A slow-growing occupation can still be a decent entry point if it has enough churn. A shrinking field can still provide steady opportunity if enough people are retiring. Labor markets are annoyingly good at refusing to fit on one chart.

The second test is structural friction. Does the work require a body in a place, a license, or judgment that carries real liability? Those things make full automation harder, no matter how slick the software gets. Then comes the practical question: what does it take to enter the field, and how long does that pipeline take to produce new workers?

High-paying jobs safe from AI aren’t really “safe”

The best way to think about AI-proof jobs that pay $100K is to split them into two routes. One route is demand-driven, where hiring grows because the underlying need is expanding. The other is replacement-driven, where openings appear because an older workforce is leaving. The tradeoff is simple enough: growth-driven jobs usually offer more momentum, while replacement-driven jobs can offer pay and stability without much expansion. But they may also be narrower, more concentrated, and harder to enter.

The BLS fastest-growing occupations list for 2024 to 2034 shows the first route clearly (BLS Fastest Growing Occupations, August 2025). The list includes nurse practitioners at 40 percent growth, data scientists at 34 percent, information security analysts at 29 percent, medical and health services managers at 23 percent, actuaries at 22 percent, operations research analysts at 21 percent, physician assistants at 20 percent, and computer and information research scientists at 20 percent. None of that proves a job is untouched by AI. It does suggest the labor market still wants people with specialized training and hard-to-copy judgment.

Healthcare is the clearest example. Nurse practitioners, physician assistants, and medical and health services managers all sit in parts of the economy where demand is tied to in-person care, regulation, and an aging population. AI may help sort information. It does not walk into an exam room and take responsibility for a diagnosis. That small distinction turns out to matter a great deal.

The education and credentialing side is part of the moat too. Nurse practitioners and physician assistants typically require a master’s degree, national certification, and state licensure, which varies by jurisdiction. Medical and health services managers usually need a bachelor’s degree, though many positions expect more. Those requirements slow supply. They also make these careers less accessible in the short run, which is the sort of bureaucratic detail that occasionally keeps a wage premium alive.

The second growth-driven cluster is more technical, and a little less romantic. Data scientists, information security analysts, actuaries, operations research analysts, and computer and information research scientists all appear on the BLS fastest-growing list (BLS Fastest Growing Occupations, August 2025). They are not “AI-proof” because they hide from software. They are durable because they sit close to the systems that organizations are trying to build, test, secure, or model.

That distinction matters. Information security analysts are growing fast because digital systems need defense, and the attack surface keeps changing. Actuaries work in an environment where probabilistic judgment and regulatory accountability still have to land on a human desk. Operations research analysts and computer and information research scientists are tied to modeling, optimization, and technical research, which means AI is more likely to become part of their toolkit than a replacement. That is an analysis, not a prophecy, but the direction is hard to miss.

What retirements create

Not every durable job needs fast growth. Some need a steady stream of replacements. That is where many readers underestimate the labor market. A field can shrink in total headcount and still produce enough openings to make sense for a newcomer. The trick is separating real openings from the illusion of expansion.

Political scientists are the neatest example in the BLS data. The median annual wage for political scientists was $139,380 in May 2024 (BLS OOH, August 2025). Employment is projected to decline 3 percent from 2024 to 2034, yet about 500 openings are expected each year on average over the decade (BLS OOH, August 2025). BLS says those openings are expected to result from the need to replace workers who transfer to other occupations or exit the labor force, such as to retire.

That last clause does a lot of work. It does not mean every opening is a retirement slot, and the source does not say so. It means replacement demand is doing the lifting. For a job seeker, that is still useful information. Five hundred openings a year in a field with about 6,500 jobs is not a flood, but it is not nothing either (BLS OOH, August 2025).

The catch is that replacement demand is a different kind of opportunity from growth. Political scientists are concentrated in a few employer types, and that concentration shapes the risk. Federal government, excluding postal service, employs 67 percent of political scientists. Professional, scientific, and technical services account for 15 percent, and educational services for 6 percent (BLS OOH, August 2025). That concentration can support strong pay, but it also makes the field sensitive to policy shifts and budget decisions. A small, elite market can be sturdy and brittle at the same time. Labor markets contain multitudes, mostly inconvenient ones.

Political scientists also show why the route into a job matters as much as the wage. BLS says the occupation typically requires a master’s degree in political science, public administration, or a related field (BLS OOH, August 2025). That is a meaningful barrier, especially in a field with limited total openings. The pay is attractive. The pipeline is not casual.

The comparison that actually helps

Growth-driven careers and replacement-driven careers look similar on paper. Both can pay well. Both can have low vulnerability to automation. Both can offer solid openings. The difference is in the mechanics.

Growth-driven jobs, especially in healthcare and technical specialties, tend to have stronger long-term demand and clearer expansion. The tradeoff is that they often require more training, more credentialing, and more time before the payoff arrives. Replacement-driven jobs can be easier to read through BLS data because the openings are explicit, but they often sit in smaller, more concentrated labor markets. A field like political science may be attractive on pay, yet narrow enough that one employer category dominates the picture. There is no free lunch. There is barely even a free snack.

For readers trying to find AI-proof jobs that pay $100K, that distinction is the one worth keeping. A career can be durable because the economy needs more of it, or because workers are leaving and somebody has to replace them. The first route usually offers scale. The second offers selectivity. Which is better depends on whether the reader wants a wider market with more upside, or a narrower one where openings are real but limited.

A practical screening method

The simplest screen for a six-figure career starts with three questions. What are the projected annual openings, not just the growth rate? What share of those openings comes from new demand versus replacing workers who leave? And does the work depend on physical presence, licensure, or judgment that is difficult to automate cleanly?

Two more questions sharpen the answer. What does entry require in time, cost, and credentials? And how concentrated is the employer base? A career with strong pay but a long training runway is not the same as one that can be entered more quickly. A career dominated by one employer type is not the same as one spread across many industries.

On that reading, the strongest candidates in this analysis are nurse practitioners, information security analysts, actuaries, medical and health services managers, operations research analysts, and computer and information research scientists. Each is listed by BLS at $100,000 or more in median pay (BLS, August 2025). Each also sits in a part of the labor market where either growth or structural friction gives it more staying power than the average occupation.

Political scientists belong in the conversation too, but for a different reason. They show that a six-figure job can still have openings even when employment is projected to decline. That is useful, provided nobody confuses replacement demand with a broad invitation. A small market is still a small market.

The BLS Occupational Outlook Handbook profiles for each occupation include projected growth, annual openings, and typical education requirements in one place (BLS OOH, August 2025). That makes the handbook a better filter than any headline that promises AI-proof work. The headline may be catchy. The data is what keeps a career from becoming an expensive surprise.