An overview of AI in the Swiss Public Sector
In December 2022, the Federal Government published its first ever Federal Data Science Strategy which focuses on “human-centric and trustworthy data science” that fosters “understanding and trust by the administration and the general public in data driven decision making”. Next to building awareness and competence, the strategy also focuses on increasing “technical accessibility and availability”, with infrastructure initiatives such as e.g. Renku playing a key role here. As an open-source knowledge infrastructure for collaborative and reproducible data science, Renku is connecting people, data and insights.
On a Federal level, the Federal Office for statistics BFS is paving the way with its Data Science Competence Center and affiliated networks such as the Community of Practice for Data Science and AI for Public Good.
Also on a Cantonal level, there are a number of initiatives. Among these, the Canton of Zurich has recently launched an innovation sandbox where researchers, private and public partners can collaborate on AI cases and which provides access to data sets on regulation and public data.
Let’s zoom in on two use cases where AI is already today applied in the Swiss Public Sector and how it specifically adds value to citizens.
Where is the growth potential for AI in Swiss public administration?
With integration of AI solutions and large language models become increasingly simple, we are expecting a significant uptake in the usage of AI solutions across all sectors. The Public Sector is no exception to this, though there are some specific considerations.
In public administration in Switzerland, we see particular growth potential for the application of AI in the areas of healthcare, social security, and transportation, both because of its relevance to the Swiss system and in light of the global use cases Deloitte has helped governments in other countries implement.
Ethics and Governance
With the rise and increasing adoption of AI, there is increasing calls to address the ethical and legal implications, especially when public organizations and governments are involved. Ethical guidelines are required for both data sourcing but also in terms of model development standards to ensure unbiased, explainable results that ensure users can place trust in AI applications and to actively prevent abuse (e.g. misinformation / deepfakes). In the Swiss context, guidelines for practitioners and users in the Swiss public administration are key and ought to be in line with the Swiss Data Protection act, with GDPR and developed in close alignment with civil society. Also, the discussion about the European AI Act should be actively followed by the relevant actors in Switzerland, both at the legislative level and for implementing bodies / users in public administration.
In 2020, the Federal Council has provided guidelines for the use of AI in public administration, focusing on transparency, explainability, and robustness. The Federal Council further clearly stated that the responsibility and liability needs to be clarified and cannot be delegated.
Canton Zurich in February 2021 published a study on ethical and legal considerations of the use of AI in the public administration Künstliche Intelligenz in der Verwaltung braucht klare Leitlinien | Kanton Zürich (zh.ch).
Conclusion
To sum it up, it can be said that with the Federal Data Science Strategy and some impressive first use cases on the way, the Swiss Public sector – and Swiss taxpayers - are starting to reap the benefits of applying AI for the public good. However, the international examples show that more can be done, and hopefully can serve as inspiration for how AI can support to relieve the burden on public personnel, thereby allowing them to focus on essential tasks and, in particular, counteract the shortage of skilled workers. Needless to say, the big requirement will be to ensure AI is applied within clearly defined guidelines and with high ethical standards to ensure it is trustworthy.
