Switzerland faces a transformation that is more profound than the industrial revolution and faster than digitalisation. Artificial intelligence will fundamentally change the labour market, the social insurance system, democracy and society's self-understanding over the coming two decades. The question is not whether, but how fast -- and whether policymakers act in time.
This chapter analyses the central risks for Switzerland: economic, societal, democratic and geopolitical. The analysis relies on publicly available data from the Federal Statistical Office (FSO), KOF ETH Zurich, McKinsey and other sources. The figures are verifiable. The conclusions are sober.
The future of AI can be described in two basic scenarios that begin with the same technology but end in entirely different worlds.
AI fills the demographic gap, takes over routine work, and an AI value creation levy closes the fiscal gap. A basic income secures livelihoods. Humanity is freed -- not free from work in the sense of laziness, but free for activities beyond mere survival. ETH and EPFL become global beacons of AI research. Direct democracy is strengthened because citizens can access AI-powered impact assessments before every vote. [1]
Humanity makes no decisions. Not the wrong ones -- none at all. It lets it happen. Gradually. Comfortably. Applauding.
AI takes over one activity after another. Each time it is celebrated as progress. The machines make decisions -- first recommendations, then suggestions, then rulings. The transition is seamless. No one remembers the moment when a human last made an important decision, because there was no such moment. It was a process, not an event.
Arthur C. Clarke formulated his third law in Profiles of the Future: "Any sufficiently advanced technology is indistinguishable from magic." [2] The point lies not in the past but in the present. We are approaching a point where AI does things we can no longer understand. Researchers call this "emergent capabilities" -- abilities that suddenly appear at a certain model size without having been programmed.
The Scottish writer Iain Banks anticipated this future in his Culture novel series: a civilisation governed by superintelligent AI beings, the "Minds". The Minds are benevolent. They manage resources, plan cities, settle conflicts. Humans live in abundance and freedom. Banks himself regarded the Culture as the best possible society. [3] Yet the disturbing question remains: what is the difference between a paradise in which you may do anything and a prison in which you need do nothing?
That is the golden cage. It is comfortable. It is safe. And the door stands open -- but no one walks through it, because outside there is nothing left that can be managed without the machines.
| Risk Area | Core Question |
|---|---|
| Labour Market | How many jobs disappear -- and what does it cost? |
| Family Budget | What does the transformation mean for a real family? |
| Disempowerment | Are we unlearning how to think? |
| Demographics | Fewer people, fewer contributors -- and AI erodes the base |
| Democracy | Can algorithms decide better than the electorate? |
| Geopolitics | Whoever controls AGI controls the world |
| Worst Case | The creeping takeover through convenience |
Both scenarios begin with the same technology. The difference lies not in AI. It lies in us. In the best case, humans remain curious, uncomfortable, defiant. In the worst case, they do nothing. And that is precisely the problem.
The door of the cage stands open. Still. The question is not whether the machines will close it. The question is whether we walk through it while we still can. For the course that must now be set cannot be set by any individual. It requires laws, investments, institutions. AI does not wait for legislative terms.
[1] IMF, AI Preparedness Index 2024. Switzerland ranks 3rd of 186 countries.
[3] Banks, Iain M.: Consider Phlebas. Macmillan, 1987. First novel in the Culture series.