I keep getting one question more and more often: "Do you really think AI will take our jobs? Will unemployment hit 10%?" I decided to answer it properly. Not with scare tactics, but with numbers that surround us today and deserve attention.

What the Data Says — and It's Concerning

The IMF estimates that about 40% of jobs globally are exposed to AI disruption — and in advanced economies, that figure is even higher. Manufacturing-heavy economies are particularly vulnerable because automation spreads fastest there.

BCG and the Aspen Institute analyzed labor market impacts: generative AI will affect more than 40% of jobs across most developed economies — involving hundreds of millions of workers — by 2035. Some positions will disappear entirely, while new ones will emerge.

But here comes the uncomfortable detail: new jobs don't emerge in the same place or for the same people as those that disappear. A factory worker doesn't become a data analyst overnight. And that's the core of the problem.

Is 10% Unemployment Realistic?

Without rapid adaptation, some economists warn that unemployment in manufacturing-heavy economies could reach 8–10% within a few years, and potentially exceed that in extreme scenarios. That would be a shock not seen since the 2008 financial crisis.

The mainstream forecast is still more cautious. The IMF talks about "structural transformation, not collapse." The WEF Future of Jobs Report 2025 globally predicts that 92 million jobs will disappear, but 170 million new ones will be created — a net global gain of 78 million jobs. That sounds positive.

But — and here I'll admit my own concern — these net numbers hide a painful transition period. Transitions hurt. Especially for those who don't have the time or resources for retraining. And there are a lot of them.

The impact is already visible in specific sectors. Companies are laying off staff with direct reference to AI automation. Junior developers report that job offers are drying up. Translators, copy editors, and administrative workers are feeling the squeeze.

Who Is Most at Risk

Based on IMF, BCG, and ManpowerGroup analyses, workers at greatest risk include:

• Translators and language specialists — AI handles translations at a professional level at a fraction of the cost
• Administrative workers — data processing, document management, customer call centers
• Junior developers — code generation, testing, and debugging are increasingly automated
• Marketing and media — writing, graphics, content production
• Accounting and HR — routine reporting and candidate screening
• Banking and call centers — AI agents now handle a significant percentage of queries without human involvement

Curiously — and somewhat unsettling — this time educated white-collar workers are in the crosshairs, not just manual laborers. AI can write contracts, analyze medical images, create financial reports, respond to customers. This is a fundamentally different dynamic from previous waves of automation, which primarily affected manual work.

How to Prepare — Concretely, Not with Empty Words

I won't just say "learn to code." That's not universal advice. Here's what the data actually shows works:

1. Learn to use AI directly in your work. The biggest risk isn't "AI will take your job," it's "a person who uses AI will take your job." Try Claude, Copilot, or Gemini in everyday work tasks — even an hour a week makes a difference.
2. Develop skills AI can't replace. Empathy, negotiation, leadership, creative strategy, and managing complex projects — these are areas where machines still fall short.
3. Look into retraining and upskilling programs. Many governments and organizations now offer subsidized digital skills courses. From 2026, retraining programs increasingly focus on sectors with real demand.
4. Don't underestimate data literacy. WEF identifies it as the fastest-growing demanded skill. You don't need to be a data analyst — but basics of data analysis, working with spreadsheets, or interpreting charts are useful in almost every field.
5. Become a hybrid: domain expert + AI user. A doctor who knows AI diagnostics is more valuable than one without it. The same goes for lawyers, journalists, accountants, HR managers. The combination of domain depth and AI tools is currently the most resilient position in the job market.

The window for proactive preparation — as the trajectory of unemployment shows — is slowly closing. Nearly every working person will need to adapt their skill set significantly in the coming years.

My Conclusion — and an Open Question

Will unemployment exceed 10% in major economies by end of 2027? Honestly: I don't know. And nobody knows for certain. But I know that waiting and hoping is not a strategy. The data shows a clear direction. Going from historically low unemployment to near-doubled rates in just one year is not normal. And it's not just a cyclical blip — it's structural change.

What worries me more than the numbers: are we as a society doing enough to ensure retraining isn't just a privilege for those who have the time, money, and access to education? Because if not, we're losing people, not just jobs.