Sweden AI Strategy Report

In May 2018, the Swedish Government released its national AI strategy National approach for artificial intelligence (Sweden, 2018). This strategy points out the general direction for AI in Sweden in order to create a basis for future policy actions and priorities. In this sense, this strategy serves as a reference for the government to outline forthcoming policy initiatives with the aim to strengthen Sweden’s welfare and competitiveness by means of AI. To this purpose, the Swedish strategy focuses on the following priority areas:

• Education and training;
• Research;
• Innovation and use;
• Framework and infrastructure.

Prior to the release of the strategy, Vinnova – Sweden’s innovation agency – published an extensive policy report outlining the opportunities and challenges of AI in Sweden, and Sweden’s capabilities to embrace the full potential of AI (with concrete examples of ongoing AI projects).

This section presents the policy recommendations of Sweden’s AI strategy. Where possible, it aims to incorporate new policy initiatives that have been rolled out since the launch of the strategy in May 2018.

In terms of funding, Vinnova funded AI projects for SEK 675 million (approx. EUR 67.5 million) in 2020. The total sum for AI projects that Vinnova helped fund was SEK 1,350 billion (approx. EUR 135 million), 50% of this could be private founding or funding from other national programs. In the national budget for innovation and research until 2024 at least SEK 550 million (approx. EUR 55 million) has been assigned to research and innovation in digital technologies and AI and its use and impact on society.

Sweden AI Policies on OECD.AI dashboard

The Swedish Government addresses the need of formal education and training as well as lifelong learning in AI by means of the following policy recommendations:

• Education institutions need to provide a sufficient number of people with AI education and training, including continuing education for professionals. Swedish universities have started proposing bachelor’s and master’s programmes in AI fields, e.g.:
  • Master programmes on Data engineering, Machine learning and statistics, Image analysis and machine learning at Uppsala University;
  • Master programmes on Design for Creative and Immersive Technology at Stockholm University;
  • Master programmes on Machine learning, and Systems, Control and Robotics at KTH Royal Institute of Technology;
  • Master programmes on Language technology and Logic at the University of Gothenburg;
  • Three newly launched AI-related master programmes on Data science and AI, High-Performance computer systems and Physics at Chalmers University;
• Non-technical programmes should include a strong AI component to foster a broad and responsible understanding of the use of AI;
• The links between research, higher education and innovation should reinforce in the field of AI;
• Continued and further education for professionals should be ensured (e.g. by offering MOOCs such as the Elements of AI course).

The following initiatives will serve to evaluate changes in skill needs on the labour market due to AI developments, and to get ready for them:

• The Agency for Economic and regional Growth together with the Swedish Higher Education Authority analyse the evolution of the supply of competencies in advanced digital technologies in both short and long term, and they make recommendation about it. The work aims to improve the supply of such competencies as well as improve statistics and forecasts of future needs in the Swedish labour market;
• A pilot project to identify the need for AI skills in companies and organisations from the southern region to better use new smart technologies based on AI. The aim is to develop AI courses depending on the needs that emerge.

Lastly, the platform AI Competence for Sweden lists university courses for working professionals in AI-related fields. The goal of this initiative is to improve the competence in AI of both academia and industry.

To transform innovative ideas into market products and services, the Swedish strategy emphasizes the need for:

• A strong basic and applied research environment in AI;
• A strong relation with leading international AI research environments;
• Pilot projects, testbeds and environments for development of AI applications;
• Efforts to prevent and manage risk associated with AI.

Some policy initiatives to address these needs are:

• AI Sweden: an ambitious holistic programme to foster the development of AI applications in Sweden. With funding from Vinnova, among others, it is organised as a national centre for applied AI research and innovation with almost 70 partners from the industrial and public sectors, research institutions, and the academic world. AI Sweden initiated and runs a number of projects supporting AI research and innovations. It also provides a platform for collaborations, and technology/data infrastructure;
• AI-related innovation projects financed through Vinnova. On 1 April 2021, a total of 256 ongoing projects matched the keywords “artificial intelligence” and “AI” in Vinnova’s project database;
• Startup AI activities: Vinnova also funds SMEs and public organisations when they start the first innovation project in AI. In a recent call for projects’ proposals participants could apply for a maximum of SEK 500,000 (approx. EUR 48,800);
• Vinnova is the national coordinator to strengthen testbeds and demonstration activities in Sweden. To this purpose it disseminates information on hundreds ongoing testbeds in Sweden, among others in the field of AI.

In terms of networking, the Swedish Government is setting up policies to:

• Foster strong collaborations and partnerships among business, the public sector and research in AI;
• Develop collaboration and partnerships on the use of AI applications with other countries.

The AI Sweden programme for instance envisages nodes and co-location areas across Sweden to boost and to enable collaborations and partnerships for AI innovation.

Another example is the Analytic Imaging Diagnostic Arena (AIDA), which is a Swedish arena for research and innovation on analytic image-based diagnostics. AIDA is a cross-disciplinary and cross-sectoral collaboration aiming to largescale usefulness of AI in healthcare.

Lastly, the AI agenda is a collaborative, stakeholder-driven, approach to suggest and formulate a national roadmap to accelerate the AI uptake in Sweden. This work is stakeholder-driven, bringing together academia, research institutes, public and private sector representatives as well as other organisations. There are six working groups looking into different aspects of AI such as: infrastructure, industry, public sector, education and more. The work is coordinated by the RISE.

In terms of international attractiveness the Swedish strategy (p. 5) claims that “if Sweden can strengthen policy conditions across all policy areas, it will be well placed to offer an internationally attractive working environment for business, researchers and others interested in AI research, development and use”.

To foster AI regulation, the Government has identified the need to:

• Develop rules, standards, norms and ethical principles for an ethical and sustainable AI and its use;
• Push for Swedish and international standards and regulations that promote a risk-free use of AI.

With regard to ethical and sustainable AI, it is important to develop ethical guidelines to ensure a transparent, explainable, and non-discriminatory development of AI. This is particularly important in systems that closely interact with the physical world, such as self-driving vehicles or AI applications in health care. To this purpose, the Swedish Government established the Committee for Technological Innovation and Ethics (KOMET) in August 2018. The Committee aims to release annual interim reports.

The launch of the following initiatives means to foster the creation of ethical and sustainable AI:

• The establishment of the AI sustainability centre: a hub co-founded by companies, universities and public authorities with a specific focus on social and ethical aspects of AI;
• Academic seminars on the ethical challenges of AI in business, administration and across various sectoral areas.

The Swedish strategy recommends to establish a legislation that fosters the use of AI and prevents risks for both society and individuals. In this sense, the new legislation should safeguard privacy, ethics, trust and social values. At the time of writing this report, the new legislation was starting hence the Swedish legislation on data protection and ownership, for instance, was still largely based on the EU law.

The Government has established the Committee for Technological Innovation and Ethics to identify policy challenges, contribute to reducing uncertainty on existing regulations and accelerate the policy development on new technologies. Automation and AI are both natural areas of interest for this committee.

The Swedish strategy also suggests to frame appropriate (international) standards. In terms of standardisation, Sweden has the following organisations and bodies:

• The Swedish Institute for Standards (SIS): an international organisation specialised in national and international standards;
• The Swedish Standards Council: the principal body for all Swedish standardisation. Its task is to promote interest in standardisation and to encourage the use of standards.

The Swedish strategy emphasizes the need for a digital infrastructure to harness the opportunities that AI can provide, including both a high-quality data infrastructure and a well-developed digital and telecommunication infrastructure in terms of computer power, connectivity and network capacity. The AI Sweden programme cover both the development of the data infrastructure – by improving data quality, data availability and data sharing opportunities, and the setting up of the IT infrastructure.

The AI Sweden programme makes datasets accessible via the Data Factory. It aims to provide horizontal resources to all research partners, ensuring that data sets are available across industries and application areas in order to accelerate AI innovation and applications.

With regard to AI in the public sector, the Swedish Agency for Digital Government (DIGG) supports AI uptake and deployment in public administrations. In a policy report from the 14^th of January 2020, the agency points out that the economic gains resulting from AI could be potentially large for the Swedish public sector. The report includes a mapping exercise on how the Swedish public sector currently uses AI, and it presents suggestions to increasingly use AI in the future. The DIGG is also supporting open data policies to foster data-driven innovations and technology developments.

Climate and environment

In October 2019, the Swedish Government and academia met representatives from U.S. academia, Google, Ericsson, USAID and UN agencies UNDP and UN Global Pulse to explore the extent of AI in accelerating innovations that help reaching the UN Sustainable Development Goals (SDGs). The event marked the beginning of the AI, People & Planet collaborative research initiative. This initiative is hosted by the Beijer Institute of Ecological Economics (at the Royal Swedish Academy of Sciences), Princeton University (Princeton Institute for International and Regional Studies), and the Stockholm Resilience Centre. It aims to deepen how AI can help expand the planetary support systems – climate stability, biodiversity, and living oceans.

In April 2020, Vinnova launched the call on AI in the service of climate, in collaboration with the government agency Formas. This call aims to support projects using AI to help Sweden having no net emissions of greenhouse gases by 2045 or to adapt society to the climate changes.

COVID-19 pandemic

The Swedish Innovation Agency Vinnova is supporting businesses and public enterprises in their innovative efforts. Among other support initiatives, it is also promoting innovation projects in AI. In April 2020, the agency announced its funding support for a project on how AI can provide more effective strategies for future pandemics.

Open science initiatives boosting access and sharing of data are essential to leverage the benefits of AI. Data and algorithms constitute key elements of AI systems. To this purpose, the Swedish COVID-19 Data Portal provides information, guidelines, tools and services also from AI to support researchers to use Swedish and European infrastructures for data sharing. To further expand this initiative, SciLifeLab, the national centre for molecular biosciences in health and environmental research, is recruiting data engineers with expertise in AI to operate the Swedish national COVID-19 research data portal and reinforce open science, and data driven life science.

Finally, the Upplands-Bro municipality on the outskirts of Stockholm is an interesting use-case of the employment of AI recruitment robots to reduce the risk of contamination at the interview stage during the COVID-19 pandemic.

The Swedish national strategy on AI will be reviewed on a regular basis to assess the policy progress and to foster the development and use of AI.

Sweden is working on a strategy for secure access to open data and the use of data as a strategic resource, with respect for rules on data protection and privacy and based on the premise that data is a basic prerequisite for being able to use the potential in AI and other digital innovation. The strategy will be published in 2021 and is an important complement to the previously published national AI strategy.

Sweden (2018). National approach to artificial intelligence. Government Offices of Sweden. https://www.government.se/491fa7/contentassets/fe2ba005fb49433587574c513a837fac/national-approach-to-artificial-intelligence.pdf

Sweden on OECD.AI observatory 

Read full report