Is AI taking the programmer's job?

AI already does part of a programmer’s job. In my own work, it has taken over most of the coding. A year ago I thought a good programmer was defined by their code; not anymore. Most of the code I am responsible for is now AI-written. My job is to decide what to build and answer for making it work. On AI and work, it pays to be brutally honest.

Finland’s future will not be built by burying our heads in the sand and pretending nothing has changed. It needs skills renewal and social security that keep pace with the change.

Is AI taking the programmer’s job?

AI already takes a significant share of a programmer’s job: writing routine code, producing tests and making repetitive fixes. It does not take the whole job. People keep the decisions about what and why to build, and the responsibility for making sure the system works and is secure. The programmer’s job is changing: value is shifting from producing code to owning the whole.

I know this from my own work. I read every change before it goes live, because AI makes mistakes in a confident voice. An error in the services I build lands directly on people’s lives and opportunities — a responsibility no machine can carry. For the same reason, a poorly defined task does not improve with AI. It only produces mistakes faster.

AI takes tasks. The whole picture and the responsibility stay with people.

I have written about AI often. In spring I wrote for the business daily Kauppalehti about what developers get paid for from now on. That is a question for the employer. The question for the employee is more direct: will I have a job in five years?

How does AI affect unemployment?

AI affects unemployment first through tasks. It rarely removes a whole occupation at once, but it cuts routine work and changes what skills employers look for. The biggest risk sits in the transitions. If learning something new takes too long, and even a short job can weaken someone’s income, spells of unemployment stretch longer than necessary.

The concern is real. The Arjen katsaus survey, commissioned by LähiTapiola, a Finnish insurer, found 62% of Finns worried about unemployment in general, and 32% fearing AI will take jobs (Yle, 27 May 2026).

In workplaces, AI is for many already the next step in development rather than an upheaval, which is why the change proceeds quietly, one task at a time. As tasks change, so does who gets hired and with what skills.

With AI, transitions become more common. More people change roles mid-career, people switch jobs more often, and some work comes as short projects. The current social security system works against this: a short job can disrupt benefits, and only a fraction of every extra euro stays in your pocket, so taking short-term work or retraining does not pay. A simpler basic security would remove this trap, so transitions no longer cost people dearly.

How will juniors become seniors from now on?

A junior will still become a senior by doing real work and getting feedback on it, but the practice tasks are changing. As AI handles routine code, young people need to learn already in school how to use AI and judge its output. It still pays for employers to hire juniors, because without them there will be no seniors in ten years.

I learned by making mistakes that more experienced colleagues helped me understand. I wrote about this in June for Skrolli, a Finnish computing-culture magazine. Now AI writes the routine code juniors once learned from, so the career path has to be rebuilt.

The first step happens in school. AI-assisted programming deserves its own subject in upper secondary and vocational education, so young people can practise judging AI’s work on real projects — for example in game development or open-source communities — before their first job.

This is also about giving everyone a chance. If only people with connections or means for unpaid internships can enter the field, the pipeline narrows and Finland wastes talent. The route into the field must stay open to everyone who wants to learn, regardless of background.

What should Finland do as AI changes work?

Finland should make skills renewal a normal part of every career, and build social security that works through every stage of change in working life, where taking work always pays off. Then AI’s productivity shows up as competitiveness, not long spells of unemployment. Finland has done this before: when Nokia’s phone business declined, its experts built new companies and a strong games industry.

The skills did not disappear from Finland — they moved to a different field. The same transition now lies ahead in knowledge work, and it must succeed by design, not chance.

I propose four concrete measures:

• Lifelong learning as a right. A right to update one’s skills would be written into law, with modular study paths between vocational and higher education, so learning something new would not require a whole new degree.
• Tax deductions for staff training. Tax deductions would support companies’ staff training, so AI’s productivity gains do not benefit only the largest companies.
• Social security where work always pays. Incentive traps would be dismantled, and the first step towards simpler basic security would be a monthly basic income of 260 euros.
• Danish-style flexicurity. Employment services would be strengthened so that moving into a new job happens fast.

Finland does well when skills renew as fast as work itself changes.

AI already makes programmers more productive. When skills and social security keep pace, that productivity makes Finland a country where talented people want to work and companies find the people they need.