One third of work is already automatable
In the Netherlands, workers estimate that on average 32% of their current tasks can already be carried out by AI or robots. At the same time, 15% say they are (very) concerned that AI will take over a large part of their work within five years.
In Belgium, both the estimated replaceability and the level of concern are higher: 35% replaceability and 21% (very) concerned about job loss due to AI within five years.
Age as a dividing line: especially young people in Belgium feel the pressure
The outcomes become clearer when age is taken into account. Belgian workers under 30 estimate that 45% of their work is already automatable, while 32% of them are (very) concerned about job loss in the longer term.
Dutch workers under 30 report a lower estimate (33%) and a considerably lower level of concern (13%). Age is therefore not only a factor in the perceived impact, but also in the experienced risk profile, particularly in Belgium.
Occupations with high AI impact: text, data and customer contact
The highest AI impact is expected by Dutch workers in roles where tasks largely consist of processing information, conducting analyses and producing text — activities where automation is relatively straightforward.
Top 4 occupations with the highest estimated AI impact:
1. Communication/PR/Journalism – 48%
2. Call center/Customer service – 46%
3. Finance/Accountancy – 43%
4. Administrative/Secretarial – 42%
The underlying logic is consistent: where work consists of document processing, data analysis and reporting, the impact of AI is more quickly perceived as significant.
At the same time, the estimated impact and level of concern do not always move in parallel. In several professions, a high degree of task automation is expected without this automatically leading to the same level of worry. This points to the importance of context: confidence in employability, adaptability, or the lasting value of human skills can mitigate the perceived threat.
People-centered roles: AI mainly seen as support
At the lower end of the ranking are people-focused professions such as healthcare, education and social work. Here, AI is more often seen as a support tool rather than a replacement.
Social and community work shows the lowest level of concern about job loss (7%), which can be explained by the core of the work: empathy, guidance and relational interaction are difficult to automate. The same mechanism is visible in healthcare. AI can reduce administrative tasks and monitoring, freeing up time for direct care, while the human element remains central.
Sectors: highest impact where processes are scalable
At the sector level in the Netherlands, the highest estimated AI impact is found in:
1. Energy supply – 47%
2. Personal/other services – 45%
3. Financial sector – 43%
4. ICT, media and telecommunications – 40%
Sectors in which proximity, physical presence and human contact are structurally important rank lower: education and especially health and social care.
The Netherlands versus Belgium: similar impact, different perception
An important insight from the comparison is that Belgian workers are more concerned across sectors than Dutch workers, even when the estimated impact is similar.
Concrete examples:
- Hospitality: concern in Belgium 24%, in the Netherlands 10%
- Accountancy: comparable impact, but concern in Belgium 39% versus 23% in the Netherlands
- Agriculture: the biggest contrast, Belgium estimates 60% impact with 41% concerned, the Netherlands 39% impact with 19% concerned
The large contrast in the agricultural sector is very likely driven by differences in existing digitalization and the adoption of precision agriculture, alongside regional differences, aging populations and economic pressure that reinforce uncertainty.
In addition, a divide becomes visible in where AI is likely to have an early impact: in Belgium, the effect appears stronger in creative applications (such as design), while the Netherlands scores relatively higher in knowledge-intensive services (such as consultancy). This aligns with the idea that the adoption of generative tools and more complex AI applications develops differently by country and domain.
Implication: focus on tasks, skills and trust
The main conclusion is not that jobs are “disappearing en masse,” but that task packages are shifting, with meaningful differences in speed and perception between groups.
Organizations that want to anticipate this effectively would do well to shift the discussion: 1)from job titles to tasks, and 2)from abstract technology to practical employability.
Key questions:
- Which activities will be eased or automated?
- Which skills will become more important?
- Where is uncertainty felt on the work floor?
That last point, sentiment, functions as an early signal. Not only of technological change, but also of the need for direction, development and reassurance.