93% of Jobs Will Still Exist in 2035. That Doesn't Mean You're Safe.

The Question Korea's Job Outlook Report Never Asked

The Korea Employment Information Service (KEIS), a government research body, published a qualitative job outlook report in 2025 with a surprising figure buried inside it. Looking at 182 major occupations in Korea and forecasting employment levels through 2035, the report found not a single occupation expected to decline. More than 93% were projected to hold steady or grow.

That number feels strange given how AI and jobs have been discussed in recent years. It was back in 2013 that Oxford researchers announced 47% of U.S. jobs were at risk of automation. More than a decade has passed since then, and since generative AI went mainstream in 2022, warnings about vanishing occupations have only grown louder. Against that backdrop, a Korean government research institute concluding that zero out of 182 occupations are shrinking comes as a surprise.

But before reading this report as grounds for relief, it's worth separating what this figure actually measured from what it didn't.

What the Forecast Says: Occupations Aren't Disappearing

The report's unit of measurement is "number of people employed." It forecasts whether the headcount in a given occupation will be higher or roughly the same in 2035 as it is today. By that measure, healthcare and caregiving occupations are expected to see steady demand growth alongside an aging population, and jobs in data analysis and AI systems operation were also placed in the growth category. Demand for content planning and creative roles, as well as counseling and coaching occupations, was likewise projected to rise.

The occupations forecast to grow share something in common: they involve stretches of work that require judgment rather than repetition, call for reading context in another person, and include trust-building that machines struggle to replicate easily. By contrast, occupations centered on standardized document processing or simple data entry were forecast to be more directly exposed to disruption. Two pressures — an aging population and technological change — are reshaping the labor market's terrain, and the direction of that change looks less like "less human work" and more like "a shift in what kind of work humans do."

It's also worth noting the scope of the survey: it covers 182 occupations that can be cleanly classified as "major domestic jobs." Forms of work that resist clean occupational classification — freelancers, sole proprietors, platform-based labor — aren't captured in this forecast data at all. With the number of sole proprietors in Korea now topping 6 million and the freelance share of the workforce steadily rising, it matters to understand just how much ground this report actually covers.

What "93% Will Hold Steady" Doesn't Say

There's a counterargument to reading this report as grounds for optimism.

An occupation surviving and the nature of the work an individual does within that occupation staying the same are two different claims. Even if the job title "tax accountant" still exists in 2035, that doesn't mean a person working as a tax accountant today will keep working the same way, for the same pay. If more people enter the market who can use AI tools to do the same work three times faster, the number of new entrants an occupation can absorb may shrink even while total headcount holds steady. Total employment staying flat and the difficulty of breaking into that field rising can both be true at once — the seats exist, but how crowded the competition around them gets is a separate question.

A 2023 OECD report contained an uncomfortable number of its own: a large share of the occupations most exposed to AI turned out to be high-skill, high-wage jobs. Contrary to the conventional assumption that repetitive, routine work gets automated first, a substantial portion of the tasks performed in professional fields — legal review, document drafting, data-driven analysis — has already reached a level generative AI can handle. That suggests employment headcounts can stay flat even as the composition of work inside professional occupations shifts rapidly.

AI technology has kept advancing quickly in the time since this forecast was published in 2025. A qualitative outlook report reflects the technology level and industry structure at the moment it was surveyed. By 2027 or 2029, the picture could look very different, and reading this report's conclusion as a blanket safety guarantee for the full decade ahead stretches the data further than it can support.

The fact that an occupation exists doesn't guarantee the position of any individual within it. Read that way, the 93% figure is more accurately treated as a signal to check your own footing than as grounds for relief.

What Sole Proprietors and Solo PMs Should Actually Do With This Data

For anyone looking to turn this report into action, the first step is narrowing the question.

Rather than asking "will my occupation still exist in 2035," ask "will this occupation still exist done the way I do it now." The job title "content director" may well still be around in 2035. But whether the work of spending 40 hours a week building proposals and reviewing drafts survives in that same form is a separate question. If more people enter the market who can produce the same output using AI tools in 10 hours a week, you need to first figure out where the other 30 hours can create value.

One practical starting point is actually laying out a full list of the tasks you currently handle. Go through each item and sort it: which tasks could be replaced by AI tools or outsourcing, and which ones can't be reproduced at the same quality without your contextual understanding and relationship capital. If the latter category turns out smaller than expected, your negotiating position can shift even while your job title stays intact. One way to test this is to pick a task you currently do, try producing it with AI tools in half the time, and see whether the quality holds up. If the result comes out similar, it's worth reconsidering how much of that task's value was really coming from you.

People who've spent a long time in one field tend to run into the same question when they check in on themselves: "What did I always assume I was good at, and does that still hold value without AI tools?" That question gets easier to postpone the longer a career runs and the more stable an occupation looks. Among people who've documented how they prepared for retirement, one of the most common regrets that shows up is: "I never once properly checked what I was actually good at while I was working." The forecast that jobs aren't disappearing can, ironically, become a reason to keep putting that check off.

Figuring out right now which of your skills hold up without AI isn't a task born of panic. It's the most practical preparation available precisely during a period when job security looks assured.

The Korea Employment Information Service's report says that ten years from now, none of 182 occupations is expected to see its headcount shrink. Which people hold which positions within those occupations, and under what conditions they work, was never a question this survey set out to answer. The 93% figure only becomes useful once it's treated as a starting point for that check, rather than an answer to it.