On 6 October, DNB announced plans to reduce its Technology & Services workforce by about 400 full-time equivalents. The bank cites AI agents and simpler processes. According to DNB, tasks such as customer checks and coding require less manual work.
Eight days later came the news that Rema 1000 plans to cut 130 office and support positions. Its stated reasons concern costs and competitiveness. E24’s report establishes no link to AI, and the cuts exclude stores and distribution.
DNB’s decision nevertheless shows what Norwegian businesses need to prepare for: more work can be done without increasing headcount. For a company, that can improve profitability. For society, the benefit also depends on what happens to the people whose work is no longer needed. Do they find another job, or remain outside employment?
Norway needs workers, while many people need work
Statistics Norway’s Labour Force Survey (AKU) reports 2.93 million employed people and 136,000 unemployed in August 2026. The employment rate was 69.9% of the population aged 15 to 74. Unemployment was 4.5% of the labour force, meaning people working or actively seeking work. These are trend estimates and may be revised.
At the same time, NAV’s 2026 employer survey estimates a shortage of 34,000 workers. Demand is particularly strong in health and care and occupations requiring vocational qualifications.
Norway could therefore have fewer office jobs and still face labour shortages. Someone losing a job in banking or administration will not necessarily have the qualifications or opportunity to fill a healthcare vacancy. Moving into new work requires education, time and employers who actually hire. General advice to “learn AI” cannot solve this alone.
Long periods outside work can increase welfare costs
NAV paid NOK 676 billion in benefits and pensions in 2025, including payments abroad. About half was old-age pensions. The fact that 2.8 million residents received money from NAV during the year therefore does not mean that as many people were out of work.
Among payments to residents of Norway, NOK 131 billion went to disability benefit, NOK 69 billion to sickness benefit, NOK 51 billion to work assessment allowance and NOK 15 billion to unemployment benefit. Municipal financial social assistance is a separate scheme: NOK 12.4 billion was paid to 165,800 recipients in 2025 as grants and loans.
These figures describe the welfare system before we can measure AI’s overall effect on the labour market. The economic risk arises if more people lose earnings and take a long time to return to work. Spending on income support can then rise while tax revenue from employment falls. Higher productivity can offset this, but that depends on where value is created, who receives the income and what can be taxed in Norway.
Public services could do more with their existing staff
DFØ reports about 181,000 state employees in 2025. This is a narrow measure that excludes municipalities, health enterprises and state-owned companies.
Across central and local government, employment was about 917,000 in the first quarter of 2026. That represents roughly 30% of employment measured on the same basis. These national-accounts figures, reproduced in an answer to Parliament, include health enterprises but exclude commercial publicly owned companies.
This is a substantial opportunity for AI: less time spent on documentation and administration could mean more time for patients, students and processing cases. The benefit could be shorter queues and better services, even with unchanged staffing. Public-sector leaders should specify how freed-up time will be used and measure whether services improve. An estimate of hours saved is not, by itself, a budget saving.
What could happen by 2031?
SSB’s September 2026 forecast provides a measured starting point: moderate growth in mainland Norway, more employed people and slightly lower unemployment through 2029.
| SSB forecast | 2026 | 2027 | 2028 | 2029 |
|---|---|---|---|---|
| Annual real mainland GDP growth | 1.4% | 2.1% | 1.5% | 1.9% |
| Annual growth in employed people | 0.5% | 0.6% | 0.6% | 0.5% |
| Unemployment rate, AKU | 4.6% | 4.4% | 4.3% | 4.2% |
The forecast does not estimate AI’s contribution separately and ends before 2031. We therefore lack a basis for putting a precise percentage on AI’s effect on Norwegian employment five years from now. In my assessment, three developments will be particularly important:
Fewer new hires could be the first sign. Businesses may stop replacing departures, reduce their use of consultants and recruit fewer junior employees. Entry into working life then becomes harder without necessarily producing large rounds of redundancies. Employers must also ensure that newcomers get the assignments they need to develop professional judgement.
Small businesses could gain capacity. A small exporter could, for example, use AI for bids, translations and customer support, serving more customers with the same workforce. If lower costs lead to new products, higher sales and investment, they can create jobs. If gains are mainly taken as cuts, the effect will be different. Competition can also push prices down, leaving an individual firm with less of the saving.
The gap between workers could widen. Experienced professionals may handle more work, while people performing routine tasks need to find another role. Some will need further education. Others need an employer willing to give them a chance in a different job. Access to those opportunities will influence who benefits from productivity growth.
I therefore expect an uneven transition through 2031. Norway could grow economically while some occupations shrink and more people struggle to enter employment. Whether the employment rate stays around today’s 70% will depend partly on how quickly new activity and jobs emerge. We should follow that alongside the duration of unemployment and the recruitment of recent graduates.
Superintelligence could make the transition much harder
Superintelligence, understood as systems that outperform humans across most intellectual work, is a more uncertain scenario. There is no sound basis for assigning it an arrival date in a five-year Norwegian forecast.
If such systems become reliable and affordable, new jobs could themselves be automated before people finish retraining. Advice to pursue more education would then be insufficient on its own. Norway needs to be able to discuss how income, ownership, taxation and working time should be shared if demand for human labour falls sharply. At the same time, faster research could deliver major advances in medicine, energy and industry.
For boards, this uncertainty means plans must remain adaptable. Follow what AI actually achieves in the organisation’s workflows, its cost after review and errors, and dependence on suppliers. Planning for more capable AI should rest on those observations.
Budget for the transition
Norwegian businesses should use AI where it delivers demonstrable value. Waiting can also cost customers and competitiveness. But investment decisions should cover more than technology and expected payroll savings.
Management should specify which tasks will disappear, which will grow and who could move into other work. Involve employees and their representatives early. Allocate time and money for training linked to real positions and needs.
Boards should request results on both economics and employment: actual gains after implementation costs, delivery quality and how many affected employees move into new work. Policymakers should likewise assess transition programmes by whether they lead to lasting employment.
Norway’s gains from AI will depend on how we use the capacity it releases. If we create new businesses, better services and new opportunities for work, productivity growth can benefit many more people than those who own the technology.
Sources checked on 14 October 2026. Assessments through 2031 are the author’s analysis. The illustration was generated using AI.