AI4SWITZERLAND
Policy

UZH Professor Karsten Donnay Says Switzerland Missed a Chance to Set Its Own AI Vision

October 21, 2026 · 6 min read · Policy

UZH Professor Karsten Donnay Says Switzerland Missed a Chance to Set Its Own AI Vision

Switzerland let slip an opportunity to signal its own vision for artificial intelligence when it chose to adopt the Council of Europe’s AI convention and regulate sector by sector, Karsten Donnay, professor of digital democracy at the University of Zurich, said in an interview published by Netzwoche on October 21. A dedicated Swiss AI bill, he argued, need not have regulated everything, but it could have shown the world which principles guide the country.

Donnay was responding to a gap between recent surveys. A future study by the association Monda Futura found that a majority of Swiss residents would rather AI played a minor role in their lives, while the IGEM Digimonitor 2026 puts regular use at about 80 percent of the population, with one person in three using it daily. He reads the first as a snapshot of unease about an unfamiliar technology and the second as everyday reality, adding that far fewer people use AI productively and that non-users tend to be under-represented in such polls.

He backed the line taken in Swiss administration that generative AI may help prepare documents while decisions affecting citizens must come from a person applying the law. More delicate, he said, is AI in politics: parliament has guidance for members, but uploading points from a session to ChatGPT would be a poor use. He sees Swiss democracy as fairly resilient so far, partly because most voters still read the official voting booklet.

What it means in practice

UZH Professor Karsten Donnay Says Switzerland Missed a Chance to Set Its Own AI Vision — contextual photo

The interview lands as Bern prepares its consultation draft for an AI law and after the Federal Council named priority areas for AI in the federal administration. His concern about workplace strain also echoes the syndicom study on rising pressure on telecom workers.

For employers, his warning is practical. Using AI so one person does the work of two can leave that person just as stressed, he said, and cutting junior hiring risks a later shortage of senior experts able to check the machines. He suggested firms may one day limit AI demands the way some now restrict after-hours email. He also cited the Diplo Foundation’s Swiss AI Trinity idea, which pairs Zurich’s innovation hub with Geneva’s international role and Swiss federalism.

What to watch: whether the consultation draft goes beyond sectoral fixes toward a broader statement of principles, and whether new surveys start measuring how people actually use AI rather than simply whether they use it.

← Back to AI4Switzerland