OpenAI released a map of Europe's AI jobs! It clearly separates the jobs that will grow from the jobs that will change

Hey there, it's Shii! Today I found a really intriguing report about the future of AI and work. It's a report where OpenAI mapped out "how will AI change jobs?" for the EU, and it's quite a satisfying read. Let me introduce it right away!

OpenAI News openai.com

What was announced?

On June 29, 2026, OpenAI published a report called "Mapping Europe's AI Workforce Opportunity." It analyzes what kind of impact AI could have on the labor market of the European Union (EU).

This report is an extension of the "AI Jobs Transition Framework," which was created for the United States in April 2026, adapted for the European labor market. Rather than simply splitting jobs into "jobs that will disappear" and "jobs that won't disappear," it's an attempt to look more closely at the type of change for each occupation.

The analysis combines Europe's official occupational classification, the "ESCO (European Skills, Competences, Qualifications and Occupations)" taxonomy, with employment statistics from Eurostat. This makes it possible to do a fairly fine-grained analysis that follows the industrial structure of each EU member state.

The story so far

There have always been plenty of stories about "AI might make jobs disappear," but many of them tended to be told through a rough dichotomy of "jobs that will disappear" versus "safe jobs." That approach failed to capture the parts where, in reality, the content of a job just changes a little, or where AI actually creates new jobs.

What's more, these estimates were often told from a US standpoint, and it was a stretch to apply them as-is to regions like the EU, where the industrial structure and labor market rules are different. Differences between countries' industries (countries with strong manufacturing, countries with strong services, and so on) were also hard to account for.

What changes

In this report, occupations are classified into the following four "transition archetypes":

• Occupations that grow alongside AI
• Occupations with a relatively high automation risk
• Occupations whose work gets reorganized (people are still central, but the way the work is done changes significantly)
• Occupations with little major change for the time being

The key point is that this lets policymakers and companies think more concretely about "what kind of support does this industry, this occupation need?" The report itself is positioned not as an employment forecast, but as a "map for planning" that shows where pressures of change and opportunities are likely to arise.

The differences between countries are interesting too. Luxembourg, Sweden, and the Netherlands are reported to have a large share of occupations that grow alongside AI, while Germany, Greece, and Italy have a large share of occupations with higher automation risk. Depending on a country's industrial structure, the way it responds going forward is likely to differ as well.

Dive Deep

Alright, from here it's serious mode! Let me look at the numbers properly.

According to the report, of the EU's employment:

• About 12% are occupations that could grow alongside AI (the type where AI lowers costs, making previously difficult projects easier to realize)
• About 14% are occupations with a relatively high automation risk in the near future
• About 27% are occupations subject to work reorganization (the type where job content and required skills will change)
• The remaining 47% are occupations with little major change for the time being

That's the breakdown. It's designed so that these four add up to exactly 100%.

What I want to highlight is the comparison with the original US framework (announced April 2026). In the US, the share of occupations deemed at high automation risk was apparently larger than in the EU, and the report positions it as "the EU has a smaller share of employment in occupations at high near-term automation risk than the US." Differences in industrial structure and the ratio of services to manufacturing seem to be behind this.

Looking at the analysis method in a bit more detail: the ESCO taxonomy is the EU-common classification system for occupations, skills, and qualifications, and by combining it with Eurostat's employment statistics, it can track the distribution of occupations at a country-by-country and industry-by-industry granularity. This is what gives it the strength of consistent comparison across EU member states.

For policy proposals, the report presents ideas like the following for public institutions and companies:

• Strengthening "monitoring capacity" to continuously track changes in the labor market
• Drawing up "readiness plans" to tailor responses to each country

OpenAI's Economic Research team, which produced the report, positions this framework as "a map for building an outlook, not a prediction," and as material for policymakers, companies, and educational institutions to think about where the pressures of change are likely to fall. So rather than taking these numbers as an absolute picture of the future, it seems best to use them as a starting point for discussion.

Wrap-up

Today I introduced OpenAI's "Mapping Europe's AI Workforce Opportunity." To recap the key points:

• Using the ESCO taxonomy and Eurostat employment data, it classifies EU occupations into four transition archetypes
• About 12% grow, about 14% face higher automation risk, about 27% are subject to reorganization, and the remaining 47% see little change for now
• Compared to the US, the EU has a smaller share of occupations at high automation risk
• Luxembourg, Sweden, and the Netherlands have many growth-type occupations, while Germany, Greece, and Italy have many automation-risk-type occupations
• It proposes that policymakers strengthen monitoring and draw up country-specific readiness plans

I really like this attitude of trying to capture the change in work with a fine-grained map like this, instead of a simple dichotomy. I'm looking forward to watching what kind of preparations each country makes from here!