AI in Switzerland doesn't simply replace jobs – it changes tasks
The world of work is changing fast. But current Swiss data argue against the simple narrative of sweeping job losses.
The wrong question is: which profession will disappear?
Artificial intelligence will change the Swiss world of work. That does not automatically mean it replaces jobs to the same extent. Anyone who asks only about professions misses the decisive point: technologies usually intervene first in individual tasks. They speed up research, drafts, translations, documentation or planning. Whether this produces fewer positions, different positions or more productive teams depends on the sector, demand, regulation and how a company organises its work.
The latest labour market data do not paint a picture of a general slump. According to the federal government, employment excluding agriculture rose by 2.0 percent year on year in the second quarter of 2026, reaching 5.698 million employed people. At the same time, the number of open positions increased. In its report on the shortage of skilled labour, the Federal Council also points out that demographic development and structural change keep demand for staff high in many areas.
AI has arrived in companies – but not to the same degree everywhere
Its spread is real, but uneven. A survey on working conditions published by SECO shows that in 2024, eleven percent of Swiss companies used machines, systems or computers with AI to carry out work tasks. That is more than the average of the EU-27, but it says little about how deeply the technology intervenes in processes. A text assistant in the office, a planning system in logistics and quality control in industry change activities in very different ways.
Recurring, rule-based and data-intensive sub-tasks are particularly easy to automate. At the same time, work remains demanding wherever responsibility, relationships, situational judgement or physical presence matter: in care, education, skilled trades, management or counselling. Even in office jobs, the activity tends to shift rather than disappear entirely. Those who check, contextualise and take responsibility for the output of AI systems take on tasks that cannot be reduced to a hit rate.
Productivity is not a question of distribution
Higher productivity can allow companies to grow despite a shortage of staff or to offer services more cheaply. But it can also increase pressure if time savings are translated exclusively into more intense work. Technology does not decide this question by itself. What matters are company rules, participation, data protection, working hours and whether employees receive further training.
SECO notes that the effects of new technologies on health and safety have not yet been researched sufficiently. This uncertainty deserves attention. AI can reduce monotonous work; but it can also strengthen control, pacing and the expectation of constant availability. A sober debate must therefore look at job quality and the distribution of power alongside the number of positions.
The transition decides how the benefits are shared
For employees, the change is concrete: those who can use new tools, scrutinise results critically and combine specialist knowledge with digital skills gain room to act. Those who get no access to training or time to learn carry the greater risk. Companies and social partners can influence this difference – for example through transparent rules of use, further training during working hours and involving teams at an early stage. Public employment services and vocational education also face the task of not reacting only after a job has been lost.
What Switzerland can influence now
Switzerland has favourable conditions: high labour force participation, a strong vocational education system and companies with experience of structural change. But this does not happen automatically. Further training must also be reachable for part-time employees, older workers and smaller companies. Just as important are clear responsibilities when AI is used in personnel decisions, performance assessments or safety-relevant processes.
The most robust statement is therefore not “AI is taking our jobs away”. It is: AI shifts tasks and can amplify or ease existing inequalities. Whether Switzerland gains more prosperity and good work from this depends less on the software than on the rules, qualifications and investments surrounding its use.
