Artificial intelligence and human concerns

The NZZ am Sonntag recently ran the headline: "AI is already causing more unemployment". At the same time, SRF reported prominently that "almost every third AI answer contains errors" and highlighted the dangers for democracy. These reports are symptomatic of the way we discuss new technologies: We can no longer see the opportunities for all the risks.Even more than that, we get caught up in the uncertainties. To illustrate this, we need to take a closer look at three risks. Firstly, immigration. Hardly any other topic is currently so hotly debated. Depending on the political coloration, either immigration itself is the problem ("too many and the wrong ones"), or the aging population, part-time work and the shortage of skilled workers, or expats and the "too attractive" location of Switzerland. But there is remarkable agreement across the political spectrum that there is a dog buried in this topic.

Then there is the fear of "job losses". The ETH study cited by the NZZ am Sonntag fuels this fear, as it focuses - as the authors themselves write - "on the potential losers of AI adoption".

And finally, according to an international study commissioned by public broadcasters, AI poses a threat to democratic opinion-forming because every third answer is incorrect and people are consuming less and less news.

These concerns are understandable, but contradictory. Either there are too many vacancies and too few skilled workers. In this case, the loss of jobs due to AI is not a loss, but an opportunity to relieve people and alleviate the shortage of skilled workers. Or we are afraid of job lossesand want to retain as many human jobs as possible. Then we can't complain about immigration at the same time. We can't have it both ways.

Similarly with AI's susceptibility to errors: those who criticize the fact that people are less well informed cannot also see a problem in the fact that they seek - and find - new sources of information. Even if they occasionally risk making mistakes in the process.

We therefore advocate a reversal of perspective. AI is primarily an opportunity, not a risk. According to a study, AI could generate an annual AI pension of 80 billion in Switzerland in the medium term - almost CHF 10,000 per person. This creates urgently needed funds to finance growing demands on health, infrastructure and security. At the same time, KI automatesworkwithout taking up an apartment in the center of Zurich or occupying the last seat on the S-Bahn. And it democratizes knowledge by facilitating access to information. Switzerland is (still) excellently positioned to benefit from this development.

The risks also lose their horror on closer inspection. Technological leaps have always changed the world of work - but rarely made it smaller. After the invention of the car, coachmen did not become unemployed, but chauffeurs did. In Switzerland, for example, over 130,000 ICT specialists will be needed by 2033. In addition, AI enables people to work more in areas where the human factor really counts: for example in care, catering and management. This is socially valuable, as most of us prefer meaningful, future-oriented tasks to monotonous routine work. And the fact that the human factor is gaining in importance offers opportunities. Also for all those who struggle to adapt to growingtechnical requirements and tend to fall by the wayside today.

Of course, KI errors - yet. The answer to this is further development, not rejection. Anyone who has been using AI for a few years can see how rapidly the quality has improved. And hand on heart: we humans also make mistakes - sometimes even deliberately, keyword "fake news".

AI harbors risks that need to be openly discussed and addressed - for example with regard to copyright for original human services such as journalism or the ability to critically question information. This ability is becoming more important for processing both artificial and human sources. But as long as unemployment remains low and we are looking for staff due to demographic change, the fear of job losses due to AI is pure projection. And justwhen we complain that debates are becoming less factual, we should see AI as an opportunity. Perhaps we should think less about machines that threaten us and more about narratives that block us. Fear is not a good advisor here.

This article was first published on 16. November 2025 in the NZZ am Sonntag.