Goals
Regulating Artificial Intelligence (AI) raises diverse and complex questions. Core challenges cover: the lack of transparency and traceability of decisions made by AI systems, risks to privacy, the danger of discrimination and risks of manipulation. Further uncertainties concern the liability for damage caused by AI systems.
The use of AI in administration as well as in the area of media and information intermediaries has particular relevance for society: In the context of administration, the state regularly acts in a sovereign manner towards its citizens. In addition, all citizens are affected by administrative actions. Whereby sensitive areas are of particular concern, for example in social insurance. The traditional media and the new information intermediaries (social media, search engines, micro-blogging services, photo and video sharing platforms) have a great influence on the formation of public opinion through the selection and presentation of content. This becomes especially evident in the run-up to votes and elections. Said phenomenon proofs increasingly problematic as information intermediaries not only influence the formation of public opinion in general, but also control the perception of content at the level of individual citizens and consumers – thus affecting them in their thinking and actions.
This project's main objective is to create a comprehensive legal framework for the use of AI in Switzerland. To this end, generally applicable legal provisions and selected sector-specific regulation concerning the public administration and the media sector (journalistic media and information intermediaries). Such norms, capable of addressing the central challenges of AI systems and preventing the occurrence of specific disadvantages for citizens are to be developed and applied. These new legal standards are presented in white papers and will be accessible to relevant parties in policy and administration, as well as to interested members of the public. The project also includes a number of measures promoting the subsequent entry of the novel norms into the legislative process.
Certain challenges can be met not only by enacting new legal norms, but also through adequate interpretation of existing legal norms. In addition to the development of new legal norms, novel interpretations and applications of existing law will be identified and presented in scientific publications. These publications are primarily aimed at expert academics, judges and administrative personnel, but may also be of interest to members of civil society.
The new legal norms and the white papers, together with the scientific publications on the interpretation and application of the existing legal norms, will form a comprehensive legal framework with which the central challenges arising from the use of AI systems can be addressed.
Focus areas
In addition to developing a generally applicable legal framework for AI in Switzerland, the project also analyses two areas in further depth: A first focus lies on the use of AI systems in public administration. Such use is subject to special rules, such as the constitutional prohibition of discrimination and fundamental procedural rights. With regard to AI applications, the right of citizens to be heard, their participation in administrative procedures and the right to a statement of reasons for state decisions raise specific questions. In this context, the prevention of discrimination and the fostering of transparency are particularly important. The legal regulation must adequately address these points without hindering the digitalisation of the administration and the use of AI.
A second focus concerns the media and information intermediaries, as the most important sources of information in democracies. Their particular impact on the formation of opinion and will is especially highlighted in connection with elections and votes. In order to take into account the new media use behaviour, the project will not only develop standards for journalistic media such as newspapers, radio or TV news, but also cover information intermediaries. Latter comprise of online services that make third-party content available to users, usually in a structured and searchable form. Information intermediaries include social media platforms such as Twitter and Facebook, as well as search engines, micro-blogging services, photo and video sharing platforms. They not only impart the content of their users ("user-generated content"), but also convey journalistic media contributions. Traditional media, on the other hand, are incorporating discussions originating from social media into their reporting. In view of this complex interaction of journalistic media and information intermediaries in opinion-forming, the project will examine whether special provisions for the use of AI systems are (also) required by Swiss law. If such need is established, research is carried out into what these should look like.
Approach
A normative framework adequately addressing the numerous issues of AI cannot be established by legal scholars alone. The project therefore follows an interdisciplinary and collaborative approach. Two interdisciplinary workshops involving colleagues from computer science, ethics, sociology, psychology as well as communication and media sciences will also ensure that the development of legal norms is based on the latest scientific knowledge.
Comprehensive solutions that are also suitable for parliamentary implementation can only be developed with the involvement of the relevant stakeholders. They include representatives of civil society, developers of AI systems and their users (companies and authorities). With regard to the focus areas, people from the public administration and from the media and information intermediaries will be consulted. In the form of multi-stakeholder dialogues, these stakeholder groups will be involved in the project from its early stages. This way, their needs and experiences can be appropriately incorporated into the development of the legal standards and white papers.
Last updated:23.08.2022
