One of the most persistent themes in discussions around Artificial Intelligence is its impact on employment. This area is fraught with contradictions, fueling a confusing public discourse that oscillates between utopian promises of enhanced productivity and dystopian fears of mass unemployment. Many workers, policymakers, and educators are understandably worried. The recent release of initial findings from the European Centre for the Development of Vocational Training (Cedefop)’s 2024 AI skills survey provides a crucial, data-driven lens through which to examine these concerns, revealing a complex and uneven landscape across Europe [1]. When combined with other recent European studies, the data suggests that the reality of AI's impact is far more nuanced than the popular narrative suggests, pointing not to a simple wave of job destruction, but to a profound and divisive transformation of work itself.

A central finding from the Cedefop survey is the emergence of a “great European AI divide.” The adoption of AI technologies is far from uniform. While over a third of workers in Northern European countries like Luxembourg, Belgium, Germany, and France report that they or their colleagues use AI at work, the figure drops to less than one in four in Poland (24%), and even lower in Greece (21%) and Spain (16%) [1]. This divergence between leading and lagging nations points to a future where the economic and social impacts of AI will be experienced very differently across the continent, creating significant challenges for cohesive European policy.

This uneven adoption feeds into the central paradox of job automation versus augmentation. The fear of job loss is real; Cedefop found that 15% of adult workers are afraid they may lose their job due to AI, a figure that rises to nearly one in five in countries like Greece, Poland, and Slovakia [1]. This anxiety is amplified by stark warnings, such as the projection from Anthropic's CEO that advanced AI models could eliminate up to half of all entry-level white-collar jobs within a few years, potentially raising unemployment by 10-20% [2]. Early evidence from the tech sector, where giants like Microsoft and Salesforce have been laying off young software developers, seems to support this grim outlook [2].

However, the broader data from European workplaces tells a more complicated story. For the majority of workers currently using AI, the primary effect is not replacement but acceleration. According to Cedefop, the main impact for 68% of these workers is the ability to perform their job tasks faster. While 30% experienced some task destruction, 41% had to take on new tasks, suggesting a shift in job roles rather than outright elimination [1]. This aligns with the findings of a comprehensive 2024-2025 survey by the European Commission’s Joint Research Centre (JRC), which found that while 30% of EU workers now use AI tools, the impact on their working conditions varies dramatically depending on the level of algorithmic management and monitoring they are subjected to [3]. The narrative of augmentation, where AI acts as a tool to enhance human capability, appears to hold more weight in practice than the narrative of substitution.

This transformation of work, however, exposes a critical challenge for the vocational education and training (VET) sector: a profound and growing skills gap. The Cedefop survey reveals a significant disconnect between the perceived need for new skills and the provision of training. A clear majority of European workers (61%) believe they will need to acquire new knowledge and skills to cope with AI's impact in the next five years. Yet a substantial portion (44%) think it is unlikely their employer will provide the necessary training to deal with AI. The numbers on actual participation are even more concerning, with only 15% of adult workers having taken part in AI-related training in the past year [1].

This training deficit is a continent-wide issue. A recent policy brief from the OECD confirms that while many countries are implementing strategies to support upskilling for AI, the current supply of training is insufficient [4]. Furthermore, the majority of existing courses focus on developing advanced, specialised AI skills, neglecting the broader need for general AI literacy across the workforce. This creates a risk of producing a small cadre of AI experts while leaving the majority of workers unprepared for the changes to their daily tasks and workflows.

Addressing this complex reality requires a more sophisticated policy response than simply promoting AI adoption. The European Policy Centre has argued for a European “AI Social Compact” to align technological progress with robust labour protections and targeted social investment, noting that current EU frameworks like the AI Act and the European Globalisation Adjustment Fund are not equipped to handle the scale of the coming job transitions [2]. The focus of many current initiatives remains on accelerating AI uptake, with less attention paid to mitigating its disruptive social consequences. As the EPC report argues, treating reskilling solely as a means to promote AI may prove counterproductive when job displacement becomes a larger reality.

For the VET community, this research provides a clear, if challenging, mandate. The impact of AI on employment is not a future problem; it is a present reality unfolding unevenly across our workplaces. The core task is not simply to train more AI specialists, but to address the wider skills needs of a workforce whose jobs are being reshaped. This means fostering what some call “hybrid intelligence”—the blend of technical, creative, and interpersonal skills that are resilient to automation [2]. It requires a focus on developing general AI literacy, enabling all workers to use AI tools effectively and critically. Most importantly, it demands that VET providers, researchers, and policymakers engage with the contradictions of the AI employment paradox, shaping a future where technology augments human potential rather than simply displacing it.

References

[1] Cedefop. (2024). Artificial Intelligence in EU Workplaces: another great divide? First insights from Cedefop’s AI skills survey, https://www.cedefop.europa.eu/files/9201_en.pdf

[2] Pozzi, F., Kuiper, E., & Valetto, P. (2025, July 25). AI’s impact on Europe’s job market: A call for a Social Compact. European Policy Centre, https://www.epc.eu/publication/ais-impact-on-europes-job-market-a-call-for-a-social-compact/

[3] European Commission Joint Research Centre. (2025, October 21). Impact of digitalisation: 30% of EU workers use AI, https://joint-research-centre.ec.europa.eu/jrc-news-and-updates/impact-digitalisation-30-eu-workers-use-ai-2025-10-21_en

[4] OECD. (2025, April 24). Bridging the AI skills gap: Is training keeping up? OECD Publishing, https://www.oecd.org/content/dam/oecd/en/publications/reports/2025/04/bridging-the-ai-skills-gap_b43c7c4a/66d0702e-en.pdf