How is Switzerland Planning to Regulate AI?

In Switzerland, considerable efforts have been undertaken to understand and guide the use of artificial intelligence (AI).

For example, the Swiss AI Report, which examines the use of AI by companies within the nation, highlights opportunities and hurdles to the adoption of the technology. Based on a survey of 92 Swiss companies, new governmental regulations were identified as critical to the effective use of AI.

This echoes the movements of policymakers across the world, who are increasingly proposing new AI regulations to promote trust, safety, and fairness. The majority of this activity is taking place in the US and the EU.

In Europe, the most significant efforts are occurring at the EU level with the AI Act, which is currently under negotiations by EU Institutions to reach a final text to pave the way for the development of standards. However, some EU Member States, such as Spain and the Netherlands, have also developed their own initiatives.

Also a member of the EU single market, Switzerland have been active in this space too. This blog post pinpoints the key elements of the nation’s AI strategy.

Digital Switzerland Strategy

As part of the 2020 Digital Switzerland Strategy, Switzerland’s Federal Council commissioned the Federal Office of Communications (OFCOM) to regularly monitor the impact of the EU's digital strategy on Switzerland through an in-depth analysis.

The 2020 strategy also outlined five main objectives of digitalisation in Switzerland:

1. Establish equal opportunities for digitalisation and protect the well-being of citizens by addressing any risks.

2. Ensure security, transparency, and trust to protect against digital misuse and unjustified prosecution.

3. Strengthen digital autonomy and self-determination to protect fundamental rights.

4. Ensure value creation, growth, and prosperity by removing barriers to market entry and trade to facilitate innovation and competition.

5. Reduce the ecological footprint and energy consumption associated with digitalisation.

While the strategy refers to digitalisation in general, it zeroes in particularly on artificial intelligence. It posits that AI could be used to limit the consumption of resources and support energy efficiency, while it also highlights the importance of ensuring that algorithmic decision systems are transparent, regulated, and respectful of social values and laws.

Guidelines on AI adoption for the federal administration

Previously, as part of the 2018 Digital Switzerland Strategy, The Federal Council established an interdepartmental working group on artificial intelligence under the supervision of the State Secretariat for Education, Research and Innovation (SERI) within the Federal Department of Economic Affairs, Education and Research (DEFR).

In 2019, the working group published a report group on the Challenges of Artificial Intelligence, which highlighted concerns such as a lack of transparency and explainability, the detection of errors, and discrimination and bias. Subsequently on 25 November 2020, the Federal Council of Switzerland published guidelines developed by the working group on AI adoption for the federal administration focused on seven key principles:

1. People at the heart – The dignity and well-being of individuals and the public must be at the forefront when developing and using AI systems to protect fundamental rights.

2. Conditions conducive to growth – Conditions conducive to strengthening value creation and sustainable development must be maintained to allow Switzerland to continue to position itself as a leading location for AI research, application, and commercial exploitation.

3. Transparency and explainability – AI decision-making processes must be identifiable and verifiable.

4. Liability – Clear liability must be established when using AI, where responsibility cannot be delegated to machines.

5. Security – AI systems must be safe, robust, and resilient by design to prevent misuse.

6. Active participation – Switzerland must actively participate in global AI governance and the development of international norms and standards in AI while also defending its own interests and values.

7. Involvement – All relevant stakeholders at the national and international levels must be included in decision-making processes on AI governance.

Artificial intelligence and international rules

More recently, on 13 April 2022, the Federal Department of Foreign Affairs published a report on AI and international rules for the Federal Council, which highlighted key issues associated with AI, including explainability, bias, protection against surveillance and manipulation, accountability and liability, and lifecycle regulation.

It also concluded that AI regulation is developing at five levels: international law, soft law, national enactments with international effect, self-binding corporate technical standards and rules, and the normative power of the factual driven by advances in technology.

In particular, the report highlighted China, the US, and the EU as key international players, all making significant headway towards legislative action to protect against AI risks.

The report also notes differences in the approach of Switzerland and other international players – Switzerland has so far taken the stance that existing laws sufficiently apply to AI so new laws are not needed, favouring a sectoral approach compared to a horizontal one.  As such, the Federal Council has called for increased efforts to converge better with global initiatives, including for the intensification of consultation with legal and technical experts, better representation of Switzerland’s position on AI in international bodies, and the promotion of exchanges with Geneva-based international standards organisations.

Closing the gap with international efforts

Switzerland is actively striving to establish itself as an AI leader, evidenced by its €10 million investment in the European Lighthouse on Secure and Safe AI. Furthermore, the Swiss National Supercomputing Centre (CSCS), in collaboration with Hewlett Packard Enterprise and NVIDIA, is jointly developing an AI-capable supercomputer.

But despite these advancements, Switzerland has not yet introduced any binding regulations or laws concerning AI use within the country. The nation’s publications to date highlight a notable disparity between Switzerland’s approach and other international approaches towards AI regulation.

Nonetheless, the Federal Council has emphatically committed to bridging this divide, ensuring the nation's values are well-represented in global discussions surrounding AI.

Holistic AI are experts in international AI regulations. Schedule a call with a member of our team to find out how we can help your organisation navigate AI’s evolving global legislative landscape.