On November 25, 2020, the Federal Council Guidelines for the use of artificial intelligence by the federal administration adopted.
The guidelines are intended to serve as a general guide for the federal administration and holders of federal administrative responsibilities when “sectoral AI strategies” are being developed, when specific regulations are being introduced or adapted in sectoral application areas, when AI systems are being developed or deployed in federal work areas, and when the federal government is collaborating on international regulatory frameworks on AI.
The following points are to be used as a guideline Principles be
- put people at the center (protection of fundamental rights, protection against discrimination, privacy by design),
- the framework conditions should ensure that the opportunities offered by AI can be exploited;
- the use of AI should be transparent, comprehensible and explainable;
- liability for AI should be clearly defined;
- AI systems must be designed to be secure, robust, and resilient;
- Switzerland should participate in the relevant international organizations and processes (e.g. UNO, OECD, ITU, UNESCO, Council of Europe and Partnership
for Peace (PfP)) remain committed to “actively shaping the global governance of AI.” - it should work to ensure that debates on the governance of AI include “all relevant stakeholders.”
As annex 1, the document contains “specific guidelines in the policy area of education, research and innovation”. Annex 2 contains some references to applicable provisions, including the revised FADP (but in extremely general terms).