Switzerland AI Strategy Report

Although Switzerland does not intend to release a national AI strategy, there is ongoing effort to promote framework conditions for research and deployment of AI in the Swiss economy and society. In December 2019, the Swiss Federal Council acknowledged the Report of the Interdepartmental Working Group on Artificial Intelligence (Switzerland, 2019a). This Working Group was established in fall 2018 by the Federal Department of Economic Affairs, Education and Research (EAER) on behalf of the Federal Council. The report provides evidence that Switzerland benefits of good conditions to encourage the deployment of AI and to tackle the numerous challenges brought about by this technology. This report analyses the existing policy initiatives in Switzerland to propose - where needed - concrete lines of actions to enable the Federal Government to fully leverage the benefits of AI. Specifically, this report provides recommendations in various policy areas and presents policy actions along 17 thematic fields.¹ Mainly, it includes:

• Improving AI-related skills and competencies at all education levels and creating lifelong learning and reskilling opportunities for the labour force;
• Fostering AI research and innovation to enhance the competitiveness of the entrepreneurial ecosystem;
• Enhancing public services through a wider adoption and use of AI applications;
• Supporting (international) networks and partnerships and ensuring the exchange of information and knowledge between all economic and institutional players;
• Establishing a regulatory and ethical framework to ensure a sustainable and trustworthy AI;
• Developing a data infrastructure to fuel AI developments;
• Reinforcing the telecommunication infrastructure, in particular with respect to cybersecurity.

In tackling these challenges for concrete policy actions, the Working Group on AI recommends close coordination and integration of policies outlined in the Digital Switzerland Strategy. Some aspects in the application of AI are addressed in the new Digital Switzerland strategy of September 2020, in particular on data use (Switzerland, 2020).

In addition, the Swiss Government adopted specific guidelines for AI in November 2020, which are intended to provide the federal administration and the agencies entrusted with administrative tasks with a general orientation framework and ensure a coherent AI policy. Regular evaluation of the application and further development of these guidelines is planned.

Switzerland AI Policies on OECD.AI dashboard

The State Secretariat for Education, Research and Innovation (SERI) has released a report on Artificial intelligence in education (Switzerland, 2019b) to enhance human capacity in AI. This report presents the policies for the adoption of AI in education, and it highlights opportunities and challenges that AI brings into the education system. Among others, it emphasises the need to develop courses to acquire the necessary skills for the deployment of AI in the primary, secondary and tertiary education systems. It includes the following policies:

• Ensuring the transfer of skills necessary for the use of AI at all levels of the education system. The State Secretariat for Education, Research and Innovation (SERI) will ensure this in close coordination with the cantons;
• Guaranteeing a transparent and responsible use of AI in education.

In addition to the need of students’ education, it is essential to guarantee that AI skills support the workforce across all sectors. Lifelong learning programmes for continuous education and training, and re-skilling and up-skilling opportunities can make this possible. These types of training have a tremendous potential and their importance will continue growing in the future.

The Working Group also recognises that the AI will change the labour market in a different way from previous technological developments. In this sense, skills and competencies of the workforce need to adapt quickly to the changing needs of the labour market. Existing measures are already in place to actively screen and monitor the skills demanded in the labour market. The State Secretariat for Economic Affairs (SECO) monitors the challenges and addresses emerging issues within the existing competencies in AI. Furthermore, in November 2017, the Federal Council decided to monitor the impact of the digital transformation on the labour market. A report will publish the results of the monitoring by the end of 2022.

The Working Group on AI highlights that Switzerland has good quality research and innovation in AI, although challenges are high.

In terms of scientific research in AI, existing activities and challenges on AI are presented in the SATW report on Artificial intelligence in science and research, prepared for the SERI. Switzerland can rely on the dynamic research environment of well-known and long-established research centres, such as the Swiss AI Lab IDSIA, the IDIAP Research Institute, the ETH Zürich Competence centres, and the Centre for Intelligent Systems at EPFL. In addition, private research initiatives and universities complement this research context. At present, the experts of the Working Group on AI suggested that the existing policy initiatives are providing the appropriate support and the Federal Government can avoid taking further policy measures. In this sense, the research capacity around AI receives support by existing policies as the Federal education, research and innovation policy 2021-2024, the Digitalisation action plan for education, research and innovation, open and competitive federal instruments, and the strategic planning of universities for 2021-2024 that identified digitalisation and AI as key priorities.

To enhance innovation in AI, the SATW report on Artificial intelligence in the industry and public administration, prepared for the SERI, presented a detailed overview of overarching challenges on AI in industry and public administration. The performance of Switzerland as for the amount and quality of AI patents, and the number of Swiss AI start-ups – reported by the Working Group on AI – reveal a strong and competitive position. As such, the Working Group concludes that the industry itself is addressing quite well the challenges of AI. However, besides self-regulation by the industry, the Working Group highlights numerous policy initiatives in priority areas such as media, mobility, healthcare, finance, agriculture and energy and climate.

Media and public

The Working Group highlights the need to govern the role of intermediaries due to the increasing use of AI in the media and to the challenges that it may bring along (e.g. fake news),. A governance report outlining concrete policy actions will be submitted to the Federal Council by the end of 2021. Other actions will tackle the monitoring of media developments and the use of AI in the media.

Autonomous mobility

The report on Autonomous mobility and artificial intelligence, prepared in 2019 for the SERI, presents governance efforts on autonomous mobility. In this respect, both the Federal Roads Office (FEDRO) and the Federal Office of Transport (FOT) are following-up the development on automated vehicles to promote data exchange (e.g. report on Provision and exchange of data for automated road driving), ensure data protection and revise the legislative framework (revision of the Road Traffic Act (SVG) and the Railways Act (EBG)).

Health care

AI is bringing many opportunities to the health system by means of data-driven medicine that can improve prevention, prediction and monitoring. The development of data-driven analytical techniques and the introduction of AI in the health care sector increase the need for data and privacy protection. Due to this, the Federal Office of Public Health (FOPH) monitors the impact of AI on medicine and healthcare also to include potential revisions to the existing legislation on the Human Research Act for data protection and privacy, and on the Federal Act on Medicinal Products and Medical Devices for the use of AI in the clinic process.

Finance

The use of AI is automating and accelerating labour-intensive processes in the financial industry too. Therefore, the need of a proper governance emerges as the use of AI in this sector expands. The Federal Department of Finance (FDF) monitors AI developments in the financial sector to fix emerging issues through proper regulatory reviews. Among others, it regulates the operational risks and it outlines the behavioural obligations to use AI methods in the financial sector.

Agriculture

In the context of agriculture, AI facilitates precision farming through image recognition and harvesting robots, among other cognitive computer technologies. The Federal Office for Agriculture (FOAG) monitors developments in agriculture on an ongoing basis. To this end, it has set up a Business Intelligence Competence Centre, which is active in the field digital data and predictive analyses. In addition, the Federal Department of Economic Affairs, Education and Research (EAER), and the FOAG launched a Charter on the digitisation of Swiss agriculture and the food industry in 2018. This Charter aims to nurture a shared awareness and promotes cooperation among relevant stakeholders.

Energy

The deployment of AI can enable significant efficiency gains in energy supply. It can support the development of renewable energies, provide energy savings and thus contribute to climate protection. Overall, it can simplify the existing complexity of energy supply operations. In this respect, the Swiss Federal Office of Energy (SFOE) monitors and tackles the AI challenges in the energy industry (see Section Error! Reference source not found. on Societal Challenges).

To foster innovations in the private sector, the creation of testbeds is recommended for cyber security and the energy sector. To increase the use of AI in cybersecurity, The National Cybersecurity Centre (NCSC), and the Federal Department of Defence, Civil Protection and Sport (DDPS), in cooperation with the Federal Department of Foreign Affairs (FDFA) and the EAER, are launching a study to evaluate the potential of a Swiss AI test centre in this field. In the energy sector, the Federal Office for Energy offers a Pilot and Demonstration Programme to promote the development and testing of new technologies, including AI-related projects.

In addition to the demonstrated tremendous potential of AI in the private sector, the use of AI is also an effective means to increase the quality and efficiency of services in the public administration. To this aim, the Federal Customs Administration (FCA), the Swiss Federal Statistical Office (FSO) and the State Secretariat for Migration (SEM), support various projects, e.g.:

• The development of a chatbot solution to reduce the costs of border crossing and the establishment of a data analytics projects to conduct risk analysis and controls in smuggling of goods. Both projects are part of the DaziT Programme, which aims to modernise and digitalise the Federal Customs Administration;
• The Arealstatistik Deep Learning – ADELE project is a deep learning application for land use and land cover classification managed by the FSO;
• The project on Automation of NOGA coding (NOGauto) proposes machine-learning methods to encode data already available at the FSO;
• The FSO project on Machine Learning – Sosi conducts data analyses on the social security system with machine-learning approaches;
• The project on Data validation with Machine Learning aims to extend and speed up data validation in the FSO by means of machine learning algorithms and at the same time to improve data quality;
• The SEM project Job algorithm for asylum seekers is a pilot test of a machine learning system to distribute asylum seekers among the cantons while optimising the labour market.

To foster similar types of projects, the Working Group on AI recommends that the federal administration encourages data exchanges and exploits the large data collections available in public administrations by means of AI-related technologies. To this purpose, a cross-administrative recording of processes and shared access to data between public departments should be envisaged. In addition, the creation of an AI competence network with a specific focus on technical aspects of the application of AI in the federal administration could facilitate the sharing of good practices.

The following policy initiatives are ongoing or recommended to promote better networking and cooperation between AI-relevant actors:

• Developing platforms to ensure dialogue and exchange of information and knowledge. The Working Group suggests that the Swiss "Plateforme Tripartite” established by the Federal Office of Communications (OFCOM) could become an interdisciplinary national competence network on AI issues;
• Strengthening collaborations between AI players by further developing hubs for digital policy debates, such as the Geneva Internet Platform (GIP). This platform has been launched by the OFCOM and the Federal Department of Foreign Affairs (FDFA) and it could act as a centre for digital governance, including AI. Strengthening the importance of Geneva and the Geneva Internet Platform as hubs for global digital and technology policy is also a main objective in the new Foreign policy strategy 2020-2023;
• Supporting the participation in the pan-European initiatives such as Horizon Europe and Digital Europe Programme, aiming to improve Europe’s competitiveness in the global digital economy through support schemes on supercomputing, Digital Innovation Hubs, and advanced digital skills among others;
• Strengthening the international cooperation for cybersecurity: the FDFA has set up the Office of the Special Envoy for Cyber Foreign and Security Policy. This targeted multilateral cooperation regularly discusses the influence of AI. At technical level, the international cooperation helps to exchange information on incident management, and at the same time the Federal Intelligence Service maintains intensive contacts with the Swiss cantons whose infrastructures are also affected by cyber-attacks.

The regulation concerns legislations and recommendations to foster AI innovations, create standard of AI adoption and application, while caring of principle of ethics and inclusion.

With respect to the development of ethical guidelines for a trustworthy, reliable, responsible and fair deployment of AI, the Swiss Government is actively involved in international discussions and committed to ensure the respect of established values and standards in the use of AI. To this purpose, it is important to guarantee principles of traceability, transparency, and inclusion (i.e. avoiding social biases and discrimination).

Towards a regulation for AI, the Working Group on AI recommends to keep the general regulatory framework to enable the development of AI in Switzerland. This framework will accommodate some clarifications and adaptations in specific thematic fields and policy areas like media, mobility, healthcare, finance, agriculture and energy and climate. The identification of these thematic areas goes along with the need to adapt sectoral regulation. However, effective regulations should target as many technologies as possible. Therefore, the Federal Government is keen to continue with a technology-neutral policy, which avoids the promotion of specific technologies and of technology-specific regulations as far as possible. While the establishment of the legal basis is ensured by a wide range of institutions, the FDFA will specifically focus on the following policies to further develop the general legal framework on AI:

• Examining the emergence of AI-specific international law and its impact on Switzerland;
• Following-up developments with regard to the visibility of AI systems in interaction with consumers;
• Monitoring developments in AI-based decision-making in the justice system (predictive justice).

In line with legislative reforms, the Working Group on AI recognises that a general improvement of standardisation and a higher interoperability would encourage AI-related research and innovation between relevant stakeholders.

The interdepartmental Working Group on AI includes the implementation of a suitable infrastructure among its challenges. The possibility to finance the infrastructure to increase the capacity in the field of AI is a challenge of technical nature. It relates to both a strong data infrastructure (i.e. data collection, data sharing practices) and a solid telecommunication infrastructure (i.e. high-speed connectivity and appropriate cyber security).

In terms of data infrastructure, following policies are mentioned:

• Supporting data exchange infrastructures in the areas of Open Access to Publication and Open Research Data. These initiatives could link up with the European electronic computer infrastructure for open science (BEAT platform) and the European Open Science Cloud (EOSC);
• Ensuring the security and protection of data through on the recently revised Federal Act on Copyright and Related Rights;
• Supporting sector-specific measures towards data collection, data sharing and data protection, e.g.:
• Releasing the Energy strategy 2050, which includes targets to building the data infrastructure to deploy AI within the energy sector (see also Section Error! Reference source not found. on Climate change). By the end of 2027, smart metering systems (so-called smart meters) will be introduced in the electricity sector. They allow digital and fine-granular data collection of electricity production/consumption;
• Establishing a digital platform (Datahub) could be envisaged to make data exchange more efficient and to make data more easily available. Standardised machine-readable interfaces (APIs) play an important role in this process as platforms and machine-readable interfaces could be at the core of the data infrastructure.

In terms of telecommunication infrastructure and associated cyber security measures, the Working Group on AI refers to a detailed report on Artificial intelligence in cybersecurity and security policy. It highlights the following initiatives:

• The National Cyber Risk Protection Strategy (NCS) presents ongoing and planned activities (2018-2022) to strengthen the protection against cyber-risks related to AI. In particular, the National Cybersecurity Centre (NCSC), formerly known as MELANI, and Governmental Computer Emergency Response Team (GovCERT) have capabilities to analyse new cyber risks related to AI;
• The National strategy for Critical Infrastructure Protection (CIP) contains 17 actions to improve the protection of critical infrastructure and thereby to ensure the availability of essential goods and services (e.g. information and communication services). This strategy also aims to grasp new AI opportunities to deliver critical services (e.g. cyber supply chain risks) and an overall better protection (e.g. AI-based monitoring and decision-making processes);
• The Cyber Defence Action Plan (CDAP) aims to systematically strengthen cyber capabilities. In addition to self-protection, the main objective is to implement the cyber aspects of the Intelligence Act and the Military Act to be able to support operators of critical infrastructure under cyber-attacks. Since 2019, the Cyber Defence Campus of Armasuisse is a platform to anticipate, detect and monitor new technologies including AI developments. The campus operates in close cooperation with both universities and economic actors;
• The Swiss Drone and Robotics Centre (SDRC) explores the opportunities and risks of combining robotics and AI for the security of Switzerland in national and bilateral projects.

Climate and environment

The objectives to cut greenhouse gas emissions in the next 30 years constitute a strong incentive for Switzerland to employ the full potential of AI to achieve environmental objectives. Issues around environmentally sound production, recycling and disposal of the necessary infrastructure and equipment for the circular economy will increasingly gain momentum.

The Working Group on AI explicitly recognises AI as key technology to meet the ecological requirements of nutrition, housing and mobility systems. To this purpose, there is an attempt to make necessary data (e.g. availability of raw materials, state of ecosystems at production sites or information in production processes) easily available and integrated into the information flows of value chains and markets. In this respect, the Federal Council released a report on Swiss Hub for Energy Data, as part of a national energy data infrastructure, which paves the way for energy strategy 2050, climate action and digital innovation. This policy report highlights that a national data infrastructure in the energy sector is essential for digitisation and innovation that enables the development and integration of renewable energy, improves energy efficiency, counters climate change and, last but not least, supports new business models.

The Federal Office for Environment ensures that environmental information is openly available within digital datasets that can possibly be used for AI applications. This institution also accompanies the environmental challenges related to AI vis-à-vis the circular economy challenges.

AI technologies can help to predict demand for energy, food or consumer goods. Furthermore, AI has the potential to reduce inefficiencies in production planning, by managing information about raw materials as much as integrating information about the production ecosystems including environmental and social aspects. Finally, AI can distribute this information to consumers in order to identify and purchase not only the cheapest but also the most environmentally friendly product for given individual consumption needs.

In sum, AI will play a key role in the energy sector, especially in the transformation from a centrally organised system to a decentralised and renewables-based one, as it optimises network planning results with forecasts of consumption and production. AI technologies can be promoted either through the targeted development of environmental applications or indirectly through the provision of large amounts of environmental data (push approach).

COVID-19 pandemic

Concerning the current pandemic of COVID-19, the Federal Office of Public Health launched the SwissCOVID app and contact tracing, this institutional page provides information about many dimensions (e.g., relevance to citizens, how to install) of the app as much as about its technical aspects (e.g., using Bluetooth and the API of Amazon and Google, using Amazon Cloud Front, using Replay attacks and AEM-tempering). Another digital initiative of the Federal Office of Public Health is the Coronavirus online check.

Furthermore, the Swiss Tropical and Public Health Institute (Swiss TPH) has a whole web page related to COVID-19 Activities at Swiss TPH and encompassing the use of AI and machine learning. The Swiss TPH is an associated institute of the University of Basel, and as a public organisation, is partially supported by the Swiss Federal Council and the Canton of Basel-Stadt. The greater part of its funding comes from competitively acquired project funds. Specifically, the following initiatives are presented:

• Mitigation strategies for communities with COVID-19 transmission in Lesotho using AI on chest x-rays and novel rapid diagnostic tests (MistraL);
• COVID-19 Outbreak Response combining E-health, Serolomics, Modelling, Artificial Intelligence and Implementation Research (CORESMA);
• Using model-based evidence to optimise medical intervention profiles and disease management strategies for COVID-19 control (MODCOVID). This approach combines mathematical models and machine learning with product development decision processes;
• Providing real-time clinical data to improve risk assessment and response, deploying an established mHealth Surveillance Outbreak Response Management and Analysis System (SORMAS).

The progress of the development and deployment of AI in Switzerland will be monitored and evaluated on a regular basis.

Switzerland (2020). Digital Switzerland Strategy. Federal Council. https://www.bakom.admin.ch/dam/bakom/en/dokumente/informationsgesellschaft/strategie/strategie_digitale_schweiz.pdf.download.pdf/Strategie-DS-2020-EN.pdf

Switzerland (2019a). Challenges of Artificial Intelligence. Report of the Interdepartmental Working Group on Artificial Intelligence to the Federal Council. Federal Department for Economic Affairs, Education and Research. State Secretariat for Education, Research and Innovation. https://ethicsandtechnology.org/wp-content/uploads/2019/12/bericht_idag_ki_d.pdf

Switzerland (2019b). Artificial Intelligence in Education (French). Federal Department for Economic Affairs, Education and Research. State Secretariat for Education, Research and Innovation. https://www.sbfi.admin.ch/dam/sbfi/de/dokumente/2019/12/k-i_bildung.pdf.download.pdf/k-i_bildung_f.pdf

1. 1) International bodies and AI, 2) Swiss intelligence of interests in the European AI (Digital Europe Programme) activities, 3) Changes in the world of work, 4) AI in industry and services, 5) AI in education, 6) Application of AI in science and research, 7) AI in cybersecurity and security policy, 8) AI, Media & Public, 9) Automated mobility and AI, 10) AI in healthcare, 11) AI in the financial sector, 12) AI in agriculture, 13) Energy, climate, environment and AI, 14) AI in administration, 15) Further development of the general legal framework on AI, 16) AI in justice, 17) AI, data and intellectual property law. 

Switzerland on OECD.AI observatory 

Read full report