Luxembourg is beginning to model how artificial intelligence (AI) might affect employment and pensions. As part of a pilot project, Statec and the General Inspectorate of Social Security (IGSS) have been working on scenarios specifically designed to analyse the impact of AI on the labour market and the pension system. Initial provisional results were presented in July 2026 at the EcoMod2026 international conference. However, the government has not provided details of the conclusions.
The information is contained in a parliamentary reply published on 3 September, following a question tabled by Alexandra Schoos (ADR). The MP was considering the implications of increasing automation for employment, wages and, by extension, the contributions that fund Luxembourg’s social security system.
The reply is addressed to the minister for labour, Marc Spautz (CSV), to the minister for health and social security,
Martine Deprez (CSV), to the minister for finance,
Gilles Roth (CSV), to the minister for the economy,
Lex Delles (DP) and the prime minister,
Luc Frieden (CSV). It is signed by Marc Spautz.
An impact that is still difficult to gauge
However, Statec’s regular analyses do not constitute a direct simulation of the impact of AI on the total wage bill or social security contribution revenue. Statec monitors its potential impact by analysing the exposure of jobs to AI technologies and by incorporating various productivity assumptions into its long-term economic projections.
Above all, these figures highlight the extent of the uncertainties. According to four scenarios, employment would rise from 522,000 people in 2025 to between 735,000 and 927,000 in 2070. There is therefore a difference of nearly 200,000 jobs between the two extreme scenarios. However, these projections do not specifically relate to artificial intelligence: they were developed as a general framework designed, amongst other things, to test different productivity trends.
Statec intends to take this work further, whilst the ministry of labour has commissioned a sector-specific study on the impact of AI in Luxembourg. This study should help to identify more clearly which sectors and which skills needs will be most affected.
When asked about a permanent monitoring body dedicated to these effects, the government refers in particular to the Labour Market and Employment Research Network (Retel) and to ‘KI-Dësch’, which brings together representatives of the State, employees, the civil service and businesses. It states that, on the basis of work currently under way, it will assess whether existing monitoring tools need to be strengthened.
No overhaul of the funding system at this stage
The issue becomes particularly sensitive when it comes to social security funding. Alexandra Schoos points out that this funding relies heavily on contributions linked to wages and employment, whilst the overall pension insurance contribution rate is already set to rise from 24% to 25.5% in 2026. She therefore asks the government whether it is considering a broader contribution base, a contribution from capital-intensive companies, or other sources of funding if, in future, an increasing share of value is generated by capital and automated production rather than by human labour.
However, the government does not, at this stage, see the need for a change on this scale. It considers that it remains difficult to make projections regarding the value added generated by AI and that the studies currently underway do not indicate a need for structural changes to the funding of the social security system.
With regard to the labour market, the government is also drawing on the Statec estimates published in 2025. These indicated that 90% of jobs could be transformed by AI, that 14% could be replaced by its use, whilst AI is expected to support 55% of the employees concerned in their work. The risk of job replacement mainly concerned administrative roles, sales and certain entry-level occupations.
The government insists, however, that this is not a forecast of job losses. The actual impact will depend, in particular, on the pace of AI adoption, its cost, the organisation of work, the skills available and the regulatory framework. Productivity gains could reduce labour requirements in certain occupations, but could also sustain demand and create new business opportunities.
Training rather than a complete overhaul
Until the situation becomes clearer, the government is focusing more on upskilling than on overhauling the social model. The national strategy ‘Accelerating Digital Sovereignty 2030’ identifies talent and skills development as one of its six key drivers to support the adoption of AI. Among the planned initiatives is the AI4Skills project, which is designed to anticipate changes in occupations and skills requirements in order to adapt training and reskilling programmes.
Adem, the Digital Learning Hub, the CNFPC and the professional chambers already offer training courses in AI, data, programming and cybersecurity. The government also intends to ensure that these schemes are accessible to older employees, those with fewer qualifications or those facing language barriers. At the same time, studies and surveys must continue in order to assess the impact of AI in Luxembourg and to determine whether current monitoring tools need to be strengthened.



