AI in communication with authorities is okay - but not unconditional
According to a study by gfs.bern, more and more people in Switzerland are accepting the use of AI in communication with public authorities - provided that data protection and human control are guaranteed. However, there are differences between language regions and age groups.
People in Switzerland are increasingly okay with authorities using AI tools for communication. These are the findings of a study conducted by Gfs.bern on behalf of the Federal Chancellery. The research institute conducted surveys to find out how the Swiss population views the use of AI in communication with authorities. The study takes into account both the personal use of AI and views on the use of this technology in public administrations, as reported by the Federal Administration.
According to the authors, the targeted use of AI for tasks such as translations, text simplification and summaries is widely accepted. When asked about their personal use of AI, 40 percent of the more than 1,000 people surveyed stated that they do not use ChatGPT and the like.
When it comes to the use of AI in communication with public authorities, the assessment varies depending on the task. For example, 80 percent of respondents support the use of AI for translations. Only 44 percent, on the other hand, would approve of AI being used for the automated writing of responses to citizens. At the same time, it is important to many that AI results are always checked by experts.
There are also regional differences in terms of trust in the authorities, according to the report. A total of 60 percent of all respondents stated that they trust the authorities. In French-speaking Switzerland, the figure is as high as 68 percent, but only 43 percent in Italian-speaking Switzerland. Younger people are significantly more likely to trust the authorities than older people - with a figure of 75 percent.
Concerns of the population and strategy of the administration
The study also shows that data protection, responsibility and transparency are central to the acceptance of AI in government communication. Accordingly, the majority of respondents demand that authorities continue to take responsibility for AI-based processes and that the results are reviewed by competent employees. Respondents also expressed concerns about dependence on international tech companies, according to the Federal Administration
Since March 21, 2025, the Federal Administration has had its own strategy for the trustworthy and competence-based use of AI systems. The associated implementation plan is currently being developed by the Federal Chancellery in close coordination with the departments. The development of AI skills is only one part of this. (Filip Sinjakovic/rja)
