Job Market 2027: Which Roles Stay, Which Disappear

The Data: What We Actually Know

The debate around AI and the labor market is dominated by two extremes: either the apocalypse is invoked (300 million jobs disappear) or concerns are dismissed (more new jobs will be created than lost). Neither is helpful. What helps is data — and there's now plenty of it.

Three things stand out:

First: Adoption speed is enormous. The share of EU enterprises using AI jumped from 13.5% to 20% within a single year — that's 6.5 percentage points, or nearly 50% growth. Among large enterprises (250+ employees), the share already stands at 55%. In the IT sector, it's 62.5%.

Second: Exposed does not mean replaced. The EU Commission's Joint Research Centre (JRC) published an updated study in March 2026 that maps 352 AI benchmarks onto 127 occupational groups. The finding: AI exposure is rising broadly across the labor market — but exposure means "support, transformation, or partial substitution." Not the binary elimination of entire professions. The strongest changes affect tasks around reading, writing, information processing, and logical reasoning.

Third: The anxiety is real. According to Cedefop, 45% of EU workers express concern about losing their job to AI. At the same time, every second worker has low or no AI-related digital skills. The skills gap is the real problem — not the technology itself.

The Vacancy Shift: Tectonic Plates Are Moving

The most important study for HR professionals comes from Cedefop. In March 2026, the European Centre for the Development of Vocational Training published an analysis of EU-wide online job postings since 2019. The result is remarkable — and contradicts the simple narratives.

This is not an abstract shift. For HR, it means concretely:

The European Perspective: Two Forces Working Against Each Other

In Europe, two opposing forces are colliding — and this makes the situation fundamentally different from the US.

Force 1: AI Adoption Is Accelerating

According to Eurostat (2025), 20% of EU enterprises already use AI — up from 13.5% just one year prior. Among large enterprises, the figure is 55%. In Germany specifically, Destatis reports 26% adoption, with the IW Cologne institute finding 37% in their own survey. The most common use cases: automating routine work, supporting complex tasks, and quality improvement. Generative AI is the fastest-growing category.

Force 2: The Baby Boomers Are Retiring

In Germany alone, according to Destatis, approximately 13.4 million workers will reach retirement age by 2039 — that's 31% of today's workforce base. Of those, 10 million are already between 55 and 64 years old today. The wave has begun.

In Austria, the workforce is projected to remain stable at around 4.78 million through 2027 (Statistik Austria). In Switzerland, unemployment stands at 3.1% (March 2026) with a forecast of 2.8% for 2027 (SECO).

The truth likely lies somewhere in between — and differs significantly by industry, region, and qualification level. What we can already say: Europe is in a different position than the US. The demographic gap acts as a buffer. But that buffer has limits — and it doesn't apply equally across all occupations.

What OpenAI Itself Says About the Consequences

In April 2026, OpenAI published a remarkable document: Industrial Policy for the Intelligence Age. Remarkable because the world's largest AI developer openly admits that its own technology will lead to massive economic disruption — and calls for political action.

The key proposals:

• Public Wealth Fund — a government investment fund, seeded by AI companies, distributing dividends to all citizens
• Robot Tax — shifting the tax burden from payroll taxes to capital gains and automated labor
• Efficiency Dividends — 32-hour work week at full pay when AI boosts productivity
• Adaptive Safety Nets — automatic benefits that kick in when displacement metrics cross thresholds
• Model Containment Playbooks — emergency protocols for AI systems that "cannot easily be recalled"

For HR professionals in Europe, these US-centric proposals aren't directly applicable. But they signal the direction: even the companies driving this disruption expect fundamental changes to the world of work — not in 10 years, but in the next 2 to 3.

Public Sentiment Is Shifting

One factor that HR departments underestimate: public perception of AI is deteriorating rapidly. According to a Quinnipiac poll (March 2026), 55% of Americans say AI does more harm than good — an increase of 11 percentage points year-over-year. Only 6% describe themselves as "very excited" about AI developments.

This directly affects HR: candidates and employees increasingly bring reservations about AI-driven processes into the workplace. Organizations that use AI in recruiting without transparent communication risk losing trust. Those that openly communicate where AI assists and where human expertise decides can turn this into a competitive advantage in employer branding.

Regulation: What Takes Effect in August 2026

Starting August 2, 2026, the high-risk obligations of the EU AI Act become fully applicable. For HR, this means: any AI system that evaluates, ranks, scores, or pre-selects candidates is subject to strict requirements — risk management systems, logging, human oversight, and transparency.

The EU Commission proposed a Digital Omnibus on AI in November 2025, featuring "targeted simplification measures" for the AI Act — including possible timeline adjustments for certain high-risk categories. Whether and when this takes effect remains open as of April 2026. The recommendation stands: don't speculate on delays — prepare now.

We covered the operational obligations for HR in detail in our article AI Development 2026: What the EU AI Act Means for HR.

What HR Can Do Now: Five Action Areas

1. Competency Mapping: Know What You Have

Before investing in upskilling, you need an inventory: which competencies exist in your team? Which are exposed to AI, which remain indispensable? A systematic competency mapping reveals not just gaps — it also uncovers untapped potential. Employees with high learning agility, problem-solving motivation, and change tolerance will become the most valuable resources in the organization over the next 18 months.

2. Adapt Your Recruiting Strategy

The Cedefop data shows: the job structure is shifting. This doesn't mean stop hiring software developers — but it means searching for different profiles. The demand is for generalists with AI competency rather than specialists whose work AI can replace. Specifically: a marketing manager who can orchestrate AI tools is more valuable than three copywriters whose output an LLM produces in minutes.

3. Upskilling With Substance, Not Symbolism

According to Cedefop, every second EU worker has low or no AI-related digital skills. A half-day "Prompt Engineering for Beginners" workshop won't change that. What helps: structured programs tied to actual work processes. Which tasks can AI take over in this specific role? What must humans do better than before?

4. Embed AI Literacy in the Organization

Since February 2, 2025, the EU AI Act requires all organizations that deploy AI to ensure "AI Literacy" among their staff (Art. 4). This is not an optional recommendation — it's a requirement for all deployers of AI systems. Starting now means avoiding a compliance problem 15 months from now.

5. Identify Potential, Not Just Manage Risk

Many organizations default to risk avoidance: which jobs will disappear, which costs will arise? Equally important: which employees have the potential to grow into new roles? Who brings the initiative, resilience, and learning motivation that become decisive in an AI-transformed workplace? These are questions that traditional HR methods cannot answer — but rigorous diagnostics can.