3 Developing and leveraging AI
Norway will focus on artificial intelligence in areas where we have competitive advantages, such as health, seas and oceans, public administration, energy and mobility. Policy instruments that stimulate investment in strong research communities, such as the centre schemes, will be important elements.
The Government wants Norwegian research communities to be attractive partners for leading AI enterprises and research communities through continued investment in basic and applied ICT research, good study programmes and competence building in AI through courses and further education programmes at all levels.
«Inside Me», Dimitry Zakharov (RU) – Photo: Ars Electronica
Norway will have advanced expertise in basic ICT research and AI research in order to understand and benefit from changes in technological developments. Norwegian communities will be attractive partners for leading business and research communities.
Norway will invest in research and development in artificial intelligence within the long-term priority areas in the Government's long-term plan for research and higher education: seas and oceans, environment, environmentally friendly energy, health, public administration and civil protection.
The EU framework programmes for research and innovation represents important arenas for cooperation and an important source of funding for Norwegian enterprises and institutions. In the future, too, the aim is for national policy instruments to stimulate participation in and qualification for any European programmes Norway chooses to join.
The EU framework programmes also open the door to cooperation with countries on other continents. Norway has entered into bilateral government agreements with selected countries to strengthen cooperation with strong research nations. The aim is to promote cooperation in priority areas, including AI.
The Government's goal is that investments in artificial intelligence within research, research-based innovation and development should be concentrated on strong research communities where cooperation between academia and industry is central, such as in the centres of excellence and the centres for research-based innovation.
Higher education programmes should correspond to needs for advanced expertise in the different sectors, both for competence in AI and in basic subjects such as statistics, mathematics and information technology.
Overviews of courses and further education programmes in artificial intelligence will be readily available. In school, digital skills, digital literacy and basic understanding of technology will be given prominence already at primary and lower secondary level. A basic introduction to artificial intelligence will be made available to the general population.
3.1 Research and higher education
There has been solid growth in research and development in ICT in recent years, with a nominal increase from NOK 8 billion in 2007 to NOK 18 billion in 2017. Industry accounts for the largest share of R&D investment in ICT, and most of this comprises development activities. More recently, however (2015–2017), the strongest growth has been in the university and university college sector. 24
Figure 3: Consumption of ICT R&D – Source: Statistics Norway and NIFO, R&D statistics, Research Council of Norway (2019): Report on science and technology indicators for Norway
In 2018 the Government allocated a total of NOK 9.9 billion to the Research Council of Norway, approximately NOK 1.4 billion of which was allocated to research, innovation and advanced applications of ICT. The Research Council of Norway has no programmes dealing specifically with AI, but AI has been one of several priorities in the IKTPLUSS initiative. Calculations from 2019 show that over 40 per cent of the Research Council of Norway's allocations to ICT was awarded to projects in AI, robotics and information management (big data). This represents almost NOK 400 million.
Between 2015 and 2019 allocations to these research areas totalled approximately NOK 1.3 billion, of which almost NOK 400 million was allocated to research infrastructure while over NOK 240 million was allocated to innovation projects in the industrial sector. AI, robotics and information management are also the ICT technology areas that grew most between 2015 and 2019 (see Figure 4). Other important ICT-related R&D areas are data protection, security and risk.
Figure 4 : R&D expenditure for different technology areas in the Research Council of Norway . Recorded expenditure of funds allocated by the Research Council of Norway 2015–2018 and budgeted figures for 2019 – Source: Research Council of Norway
The investment in AI is heavily influenced by the challenges and opportunities that lie in applications, and projects that combine good AI research with advanced applications stand a strong chance of winning the competition for funding.
The Research Council of Norway's centre schemes
The centre schemes are intended to support the best research communities and drive pioneering research and ground-breaking innovation or to strengthen priority areas. The centres for research-based innovation and the centres of excellence are two examples of such schemes. The overarching objectives of the centres for research-based innovation are to enhance innovative ability and increase value creation in the Norwegian business sector through long-term research.
BigInsight – Statistics for the knowledge economy and Sirius – Centre for Scalable Data Access are centres for research-based innovation that focus on AI-relevant technologies (big data and data analysis). The Research Council of Norway is now witnessing a trend whereby most centres for research-based innovation have one or more activities related to artificial intelligence. Sirius is affiliated with the BigMed project, to which the Research Council of Norway has allocated NOK 60 million for personalised medicine and use of large-scale data analysis in healthcare.
Norwegian AI Research Consortium (NORA)
In 2018 a consortium was set up aimed at strengthening Norwegian research and education in artificial intelligence, machine learning, robotics and related disciplines. The consortium comprises Norwegian universities and research institutions engaged in research and education in artificial intelligence: the University of Agder, the Arctic University of Norway, OsloMet, the University of Bergen, the Norwegian University of Life Sciences, Simula Research Laboratory AS, the University of Stavanger, NORCE and the University of Oslo.
Industrial PhD scheme
Under the industrial PhD scheme, companies can get funding from the Research Council of Norway to allow an employee to undertake a PhD project in an area that is relevant to the company's activities. The scheme enables companies to enhance their long-term research competence and innovation ability and to strengthen their collaborative relationship with academia.
The scheme is a good alternative for companies in need of targeted research activities but with no possibility or capacity to initiate large-scale R&D projects. Figures from the Research Council of Norway show that artificial intelligence is the area of knowledge in ICT showing the strongest growth in the industrial PhD scheme.
The Government's ambition for Norwegian AI research
The Government considers Norway to be well placed to succeed with human-friendly and trustworthy AI, and with AI in industrial applications. Norway has a technologically advanced industry with the ability and will to adopt new solutions. Although large countries such as the United States and China have resources with which neither Norway nor Europe can compete, there are areas where Norway and Europe have competitive advantages, such as certain industrial applications of AI and trustworthy AI that takes data protection and ethical considerations into account.
To ensure that Norway remains internationally competitive in areas where we have comparative advantages, it is important that AI research be integrated with investments in these areas.
As a small country, Norway does not have the capability to build knowledge and expertise to high international standards across the full spectrum of AI. But the quality and scope of our national expertise must be sufficient to exploit the technologies and innovations that emerge internationally. Another goal must be to leverage our position as a nation with a digitally advanced population and business sector in order to take the lead in applying AI, not least in industry.
Our national research and study programmes in artificial intelligence must hold a high international standard. Artificial intelligence falls under one of the five long-term priorities in the Government's long-term plan for research and higher education: enabling and industrial technologies. The Government is increasing allocations to this priority area through, among other things, the Technology Initiative escalation plan, which is one of the stated measures in the long-term plan.
It is also important that Norwegian research communities take full advantage of the opportunities offered by the EU framework programme, and that we leverage the opportunities that lie in our bilateral research and technology agreements with strong academic communities in leading research nations worldwide. Norway must be a trustworthy and attractive partner in international cooperation, networks and consortia dealing with research and education in artificial intelligence.
A key research policy objective is to stimulate private investment in research. 25 This includes stimulating private individuals and organisations to make donations to research. There are general policy instruments to support this: under the gift reinforcement programme, donations of at least NOK 3 million made to long-term, basic research can trigger a gift reinforcement from the state worth 25 per cent of the donated amount. This programme is administered by the Research Council of Norway. The Norwegian Tax Administration has a scheme under which deductions from taxable income may be granted for monetary donations made towards scientific research conducted by pre-approved organisations.
Private investments in AI research
Wallenberg AI, Autonomous Systems and Software Program (WASP)
Wallenberg AI, Autonomous Systems and Software Program (WASP) is a Swedish research institution funded by the Knut and Alice Wallenberg Foundation. WASP cooperates with Sweden's five leading ICT universities. WASP focuses on two areas in AI: the main focus area covers machine learning, deep learning and next-generation AI (explainable AI) and the other focus area is the mathematical foundations of AI.
WASP also aims to recruit outstanding international researchers and to enhance competence in Sweden and in Swedish industry by training at least 400 new PhD candidates, of which at least 100 should be industrial PhD candidates. WASP has a budget of SEK 3 billion up to and including 2026.
Norwegian Open AI Lab
The Telenor-NTNU AI Lab was established at the Norwegian University of Science and Technology (NTNU) in 2017 after Telenor donated approximately NOK 50 million. The donation was made to help boost research, innovation and education in artificial intelligence, machine learning, big data analysis and IoT in Norway. In 2018 several business partners were included in the cooperation project, and the research centre changed its name to the Norwegian Open AI Lab. The new partners – DNB, DNV GL, Equinor and Kongsberg – have donated funds towards general research at NTNU and to research where AI is a central component. The contributions from industry are donated primarily to fund researcher positions and professorships but also physical facilities and equipment that can be used by both researchers and students.
These types of donations from industry strengthen relationships between the partner companies and the university. The connection to industry forms a better platform for exploiting the opportunities that lie in artificial intelligence. This form of cooperation will, among other things, give students and researchers opportunities to work on authentic problems and datasets from industry, contributing to innovation and moving the frontier of AI research forward. Cooperation is continually being expanded to include new cooperation partners from research and industry.
Sources: www.wasp-sweden.org and Norwegian Open AI Lab
Participation in EU programmes
Norway has actively participated in the EU's framework programmes for research and innovation for many decades. The Research Council of Norway's policy instruments should be designed in such a way that they stimulate and qualify for participation in the EU programmes. Artificial intelligence and high-perfor¬mance computing are areas where we are dependent on international cooperation.
Artificial intelligence is included in several calls for proposals in Horizon 2020, and will have an even more prominent place in the next EU framework programme, Horizon Europe. Investment in innovation will also be strengthened compared with the current period.
The EU has proposed establishing a comprehensive digitalisation programme, Digital Europe Programme (DEP), for the period 2021–2027. This is the first time the EU has proposed a programme dedicated specifically to digitalisation. The programme has a proposed budget of EUR 9.2 billion divided into five main areas: high-performance computing, artificial intelligence, cyber security, digital transformation and interoperability, and advanced digital skills. DEP is a core element in the European Commission's efforts in digital transformation. The programme targets areas where individual countries alone lack the necessary resources to succeed. The resources in the programme will be used in areas where they are expected to be most effective, such as health, justice, consumer protection and public administration. The programme also aims to help small and medium-sized enterprises to adapt to changes brought about by digitalisation and artificial intelligence.
Norway has made a non-binding declaration of intent to participate in the next programme period for the research and innovation framework programme, Horizon Europe. This will give us the opportunity to participate in the strategic planning process. The Government is considering Norway’s possible participation in DEP.
Figure 5: Priority areas in DEP – Source: European Commission 2018
The Government will also facilitate participation by the business sector in the European Commission's research, innovation and development programmes. This will give access to leading knowledge communities and markets. The Research Council of Norway and Innovation Norway have policy instruments for providing guidance and encouraging Norwegian actors to participate in EU programmes.
The Research Council of Norway has taken the initiative to form a national network for artificial intelligence. The objective is to promote Norway's priorities and interests in the strategy work that takes place in the EU and to achieve greater participation by Norway in Horizon 2020 and Horizon Europe. The network will coordinate input from Norwegian research communities and give strategic advice concerning the Research Council of Norway's focus areas. An important task for the network is to organise meeting places to promote contact, cooperation and knowledge dissemination between the research communities and increased involvement by the business and public sectors. The network was established in the spring of 2019 and brings Norway's leading AI communities together.
The universities and university colleges have a high degree of autonomy. They are responsible for following up the national sectoral goals for the university and university college sector, one of which is to facilitate good access to education. The institutions have responsibility for designing the study programmes. The Government can use various means to signal which areas it wants the institutions to give priority to; for example, by allocating resources to study places earmarked for certain subject areas, such as nursing or cyber security. Since 2015 the Government has given particular priority to ICT-related programmes. Increased allocations to these programmes allow almost 1,600 more students admission to ICT-related studies every year.
Subjects that can be classified as artificial intelligence are part of several study programmes, but are most common in programmes in information technology and cybernetics/robotics. In these programmes – at both bachelor and master level – students may, to varying extents, choose courses in AI or decide to specialise in AI by both choosing AI-relevant subjects and writing their bachelor or master's thesis in that subject area. Typical subjects in these types of study programmes are: algorithms, machine learning, computer vision, deep learning and big data analysis.
Some educational institutions also offer dedicated programmes in AI, but up to now these programmes could offer relatively few study places. Since 2018, however, we see that a growing number of educational institutions are offering separate programmes in AI at both bachelor and master level. Institutions already offering such programmes are expanding the number of study places. More than 350 new study places dedicated to AI, robotics and data science are planned from 2020, compared to 2017.
More than 400 master degree candidates with a clear AI profile are estimated to have graduated every year from 2016 to 2019. These were candidates from study programmes dedicated to AI and candidates whose choice of course and thesis defined them as 'AI candidates'. 26 The majority of these students graduated from the University of Oslo and NTNU. In addition, just over 300 bachelor candidates graduate every year with study profiles in artificial intelligence. Some of them pursue master programmes in the same subject area.
Informal learning arenas: Eik idéverksted
Eik idéverksted is a makerspace for technological innovation. The makerspace is a cooperation project between the Norwegian University of Life Sciences, Inkubator Ås, the Savings Bank Foundation DNB and the SpareBank 1 Foundation. Eik idéverksted offers a range of courses, workshops, seminars and projects to students and employees at the Norwegian University of Life Sciences. The aim is to serve as an active meeting place that is open to the academic, student, local and business communities.
Many of the students who participate in Eik idéverksted learn to use AI tools at quite advanced levels, even though they do not necessarily have the same in-depth knowledge and academic background as students taking formal courses in AI. AI is becoming increasingly accessible because companies like Google and Facebook offer low-threshold tools in areas such as image processing, natural language processing, prediction, and so on.
Sources: Norwegian University of Life Sciences (NMBU), eikide.org
The number of PhD degrees in topics that can be defined as artificial intelligence is also steadily growing. Moreover, methods from artificial intelligence are increasingly used as tools by PhD candidates and researchers in other areas.
As well as seeking candidates with specific training in AI and data science, business and industry is also interested in candidates with sound knowledge in information technology, mathematics and statistics. Candidates like these will possess the necessary prerequisites for understanding and learning more about AI, either through formal education or through training in the workplace.
Development of study programmes in AI
The Faculty of Mathematics and Natural Sciences at the University of Oslo hosts a centre of excellence, the Centre for Computing in Science Education (CSSE), which makes an important contribution to further developing study programmes in AI.
The faculty, through CCSE and the Centre for Teaching and Learning in Science and Technology (KURT), offers comprehensive continuing education to schoolteachers in order to integrate programming into school subjects. This is done through the ProFag programming course for schoolteachers. What subjects are taught in schools is a key factor in raising the level of higher education in artificial intelligence, including machine learning, robotics and data science.
Source: University of Oslo
Need for interdisciplinarity
Knowledge of artificial intelligence, and related fields such as ethics and data protection associated with AI applications, will be important in study programmes oriented towards the educational sector, health, crime prevention, law and several other fields. For example, an introductory course in artificial intelligence has been established at the University of Bergen, in which students of medicine and bioengineering learn about how AI can be used in clinical practice. One of the aims is to promote interdisciplinary cooperation between physicians and engineers. The universities of Bergen and Oslo also offer courses in artificial intelligence and machine learning oriented towards social scientists.
Institutions of higher education ought to evaluate how topics with relevance to artificial intelligence can be integrated into their programmes in areas that will be affected by artificial intelligence in the coming years.
Need for knowledge about study programmes with relevance to artificial intelligence
During the work on preparing this strategy, it has become clear that official statistics on higher education are not sufficiently detailed to provide reliable information about the scope of the different study programmes needed to succeed with artificial intelligence in Norway. This knowledge base must therefore be strengthened. The EU is also working on models that can produce better knowledge on this issue, based on official statistics and existing databases. Norway is engaged in this work, and will evaluate whether new measurements can be based on these models.
The Government will
- strengthen basic ICT research through the Research Council of Norway
- complete implementation of the Technology Initiative escalation plan laid out in Report to the Storting no. 4 (2018–2019) Long-term plan for research and higher education 2019–2028
- consider Norwegian participation in relevant EU programmes in the next programme period
- take an initiative to encourage research cooperation with the private sector, inspired by the Wallenberg AI, Autonomous Systems and Software Program (WASP) in Sweden
- set clear expectations for educational institutions to design and adapt their study programmes in AI to meet anticipated needs in the labour market and to integrate AI into established study programmes where relevant
- establish a knowledge base that makes it possible to monitor trends in study places and candidates in AI
The Government wants digital skills and technology literacy to be given more prominence already at primary and lower secondary school level. Under The Curriculum Renewal, 27 natural science will be made a more exploratory and practical subject already at primary school level, with a distinct technology element that also includes programming. Programming will be introduced in more subjects; among other things, programming and computational thinking skills will be included in the new mathematics subject.
The new curricula will come into effect in 2020. It will take a long time before the children starting school in the autumn of 2020 complete their upper secondary education and be ready to start work or take further studies, perhaps in technology or artificial intelligence. In the meantime there will be a need to manage the new skill requirements expected of today's adult population in both their private and working lives.
Courses and further education programmes
An analysis from 2015 estimated a high probability that 33 per cent of jobs in Norway would be automated over the next 20 years. These figures have subsequently been modified, and a 2018 report from the Organisation for Economic Co-operation and Development (OECD) 28 estimated that only 6 per cent of jobs in Norway were at risk of automation. This figure is the lowest of all OECD member countries. It also emerged that Norwegian employees – both those with and those without higher education – were among those who receive most training in the workplace.
NAV's Horizon Scan 29 deals with the most important societal trends that will affect the future labour market. It shows that the pace of change in the labour market will accelerate. OECD estimates that every fourth job will undergo major changes. Technological developments will likely lead to many new jobs, but will just as likely lead to employees having to change jobs, change the formal relationship they have with employers, and update their skills more often. Opportunities for upskilling and reskilling – both in the workplace and through study programmes – will therefore become increasingly important. Development and use of artificial intelligence make up part of this picture.
AI for everyone: Elements of AI
Elements of AI is a series of free online courses in artificial intelligence. The course was developed in the spring of 2018 by the Finnish company Reaktor, its Norwegian sister company Feed and the University of Helsinki. The aim of the course is to teach as many as possible about what AI is, what it can and cannot be used for, and how to start using methods based on AI. The course combines theory with practical exercises, and requires no prior knowledge.
Elements of AI has quickly become extremely popular, and ranks top in several overviews of massive open online courses (MOOCs). Eighteen months after it was launched, more than 230,000 people had registered to take the course, of which 15 to 20 per cent have completed all the modules.
The course has attracted much larger proportions of female students (40 per cent) and students aged over 45 (30 per cent) than is normally the case for ICT courses. In Sweden and Finland the proportion of female students taking the course is more than 50 per cent. Since Elements of AI was launched, the University of Helsinki, which is the host university, has seen increases in the number of ordinary applicants to ICT programmes and in the proportion of female applicants.
A key factor in motivating people to complete the course is the #AIChallenge. The #AIChallenge encourages companies and public agencies to commit to supporting their employees to complete Elements of AI. By the autumn of 2019 more than 450 Finnish and Swedish organisations had taken up the challenge, among them Telia, Nokia, Systembolaget, Nordea, Bonnier, Swedish Tax Agency and Volvo.
Elements of AI will be launched in Norwegian in cooperation with NTNU in 2020.
Like elsewhere, technological developments in artificial intelligence are bringing about changes in many jobs; some work tasks will disappear, while tasks that require other skill types will emerge. Meanwhile, ever fewer jobs will require little or no formal skills. The Government considers it important that people who already have jobs be given opportunities to take courses and further education. In 2020 the Government will present a white paper on a skills reform for lifelong learning ( Lære hele livet ). The objectives of the skills reform are that no one should be left behind and that more employees should remain longer in their jobs.
Employers must ensure that their employees possess the necessary skills. This includes assessing the need for continuing and further education. Individual employees must also take responsibility for their own skills development in a labour market with ever increasing requirements. Nonetheless, some skill needs may not be met, and the public sector ought to find new ways of addressing these.
To help enterprises meet new skill needs that arise, the Government has established a range of schemes that support the development and operation of flexible further education programmes. Several programmes in digital skills have already been established as a result of cooperation between universities, university colleges or tertiary vocational colleges and the labour market. The idea is to enable people to undertake this type of further education while they are employed.
From 2020 the Government will reinforce these efforts through a new Skills Programme and through increased allocations to a competition-based scheme administered by the Norwegian Agency for International Cooperation and Quality Enhancement in Higher Education (Diku). The target group for the programmes developed with funding from these schemes are employees who need new skills due to restructuring needs created by digitalisation and the transition to a green society.
Norway has positive experiences with handling restructuring in the labour market due to various changes. 30 Norwegian employees are good at adapting, and cooperation between employee organisations, employer organisations and the authorities is good. This may be one of the reasons why Norwegian employee organisations tend to seek dialogue on how employees and employers can work together to meet the coming challenges rather than focus their concerns on the potential negative consequences of AI for jobs and work tasks.
Several large enterprises have launched their own further education programmes in AI and data analysis. For example, DNB bank has offered its employees opportunities to train as data scientists in order to meet the bank's need for competence in this area.
Norwegian State Educational Loan Fund: testing AI as a training measure
When the Norwegian State Educational Loan Fund (Lånekassen) was introducing a new collaboration platform in 2017, it wanted to find out whether an AI chatbot could make training users of the platform more effective. Lånekassen decided to use a technology supplied by the Norwegian company Boost.ai, a market leader in conversational AI. Because testing the chatbot involved all Lånekassen employees, they all gained insight into what a chatbot is and how it is trained.
After two weeks, Lånekassen's employees had performed close to 1,500 chats with the robot. The system made an important contribution to the launch of the collaboration platform, as well as illustrating how use of AI created an opportunity to involve a whole organisation. This measure is an example of successful public–private cooperation on using new technology and modern development methodology. The test project has also provided experience transfer for other public agencies such as Vy, NAV and the Norwegian Government Agency for Financial Management (DFØ).
Source: Lånekassen (2017): Sluttrapport fra konseptutreding av muligheter for effektivisering ved bruk av kunstig intelligens [Norwegian State Educational Loan Fund (2017): Final report from the concept study of the rationalisation potential of AI]
The Government is also cooperating with the employee and employer organisations on an industry programme for the municipal care sector and an industry programme for industry and construction. The costs of the industry programmes are shared among the state, which pays for developing and operating the programmes; the enterprises, which invest their own resources; and the employees, who invest some of their free time. The industry programmes will also be included in the Skills Programme in 2020, and work on the programmes will be strengthened and further developed in cooperation between the partners.
Student projects as a measure to access new skills and for recruitment
The Norwegian Tax Administration has in recent years hosted summer projects using AI, such as Pattern Recognition in Cryptocurrency and AI-assisted Tax Returns. The objective of these projects is to test new technologies. Participants are recruited from technology study programmes in AI, machine learning, programming and statistical methods, among others. The projects not only provide the Norwegian Tax Administration with an opportunity to conduct interesting pilot projects in relevant areas; they also give the agency the chance to present itself as a potential workplace for students who will be highly sought-after in the labour market once they graduate.
Source: Norwegian Tax Administration
The Government will
- make Elements of AI available in Norwegian through the Norwegian University of Science and Technology, and initiate the #AIChallenge in Norway
- consider a digital platform for continuing and further education programmes
- support development of flexible further education programmes so that universities and university colleges can apply for funding to develop programmes in AI
- prepare a strategy for digital competence in the public sector
The enquiry regarding the numbers of study places and graduated candidates in AI was sent to all the universities as well as to Østfold University College, Western Norway University of Applied Sciences, Inland Norway University of Applied Sciences, and Kristiania University College. All of them submitted reports. As well as reporting the number of study places and the number of candidates who took examinations at bachelor, master and PhD levels in AI, they also reported on new study programmes in AI and on increases in the number of study places in existing AI programmes. The institutions may have used different definitions of what constitutes an 'AI profile' or an 'AI programme' and thus which candidates were counted.