4 Enhancing innovation capacity using AI
The Government wants Norway to take a leading position in exploiting the innovation potential in applying artificial intelligence. The Government will evaluate how industrial policy instruments can best be designed to support potential value creation and use of AI in the business sector.
Public agencies ought to actively explore the potential of the technology, and increased interaction between the public sector and the business sector should promote innovation and value creation.
«Aerial biped», Azumi Maekawa (JP) og Shunji Yamanaka (JP) – Photo: Ars Electronica
Norway can take a leading position in applying artificial intelligence, particularly in areas where we already are well positioned and have strong business and research communities, such as health, oil and gas, energy, the maritime and marine industries and the public sector.
The foundation we lay in the form of access to good infrastructure, data sharing, research and competence building will also provide a good starting point for increased innovation and value creation for small but technologically advanced companies.
The largest public agencies already serve as a driving force in AI by actively exploring the potential that lies in this technology. The Government will facilitate sharing of best practice across sectors and enterprises.
Increased interaction and cooperation between the public and business sectors, and between research communities and the business sector, are vital to unlocking the innovative potential of applying artificial intelligence in the form of both established and new models of cooperation.
The public sector ought to actively explore opportunities in the market in connection with procurements, and innovative public procurements should be used wherever appropriate. To facilitate innovative solutions, agencies ought to focus on their needs rather than on specific products or services.
4.1 Industrial policy instruments
The authorities play an important role in facilitating business development, also with regard to AI. However, the business sector also has a responsibility to develop and adopt technologies that can create better, more profitable services or more efficient operations.
The underlying assumption is that companies and enterprises invest less in research and development than what is profitable for society as a whole. As part of Norway's research and innovation system, public authorities therefore facilitate innovation through grants and other schemes administered by the system of policy instruments for research and innovation. Several policy instruments currently available promote research in, and development, testing and commercialisation of, artificial intelligence.
Support for the early phase, where more emphasis is placed on research and research-based innovation, is usually provided through the Research Council of Norway.
Innovation Norway has a responsibility to promote innovation, value creation and growth in business and industry through, among other things, financial contributions and expertise. In recent years Innovation Norway has arranged the Tech City Executive Accelerator (TEA) initiative for companies oriented towards artificial intelligence and IoT. The aim is to make executives in expanding Norwegian companies aware of the opportunities for leveraging AI and IoT. TEA is based in London.
The Industrial Development Corporation of Norway (Siva) facilitates innovation through infrastructure such as business gardens, incubators and initiatives such as the Norwegian Catapult scheme. Investinor is a state-owned venture investment company whose purpose is to improve access to capital in the early-stage market (the market for investment capital for companies in the process of developing new products or processes). Export credit and guarantee schemes help Norwegian businesses enhance their competitiveness in the international market.
Innovation project: Machine learning in seismology
Earth Science Analytics AS leads an industry innovation project that receives funding via the Research Council of Norway's PETROMAKS 2 programme Machine Learning in Geoscience. The project shows how digitalisation and artificial intelligence are in the process of transforming the seismic industry.
Today geologists spend around 70 per cent of their time on seismic interpretation. Artificial intelligence could automate some parts of the interpretation process. This would make retrieving relevant information from seismic data far easier and more efficient. The technology can also be combined with tools for analysing well data. This will help the oil companies make better decisions based on large amounts of reliable data.
Source: Ministry of Petroleum and Energy
Some countries have introduced grant schemes for businesses that start up AI projects. In Sweden, for example, Vinnova has established a scheme to which companies and public enterprises can apply for grants of up to SEK 500,000 to start up their first AI project. Norway has no schemes specifically for R&D in AI; instead we have a broader scheme, SkatteFUNN, through which Norwegian companies can apply for tax deductions on R&D costs. SkatteFUNN is a rights-based scheme with simple application processes and reporting requirements.
Forskerpool [Researcher Pool] is a scheme affiliated to SkatteFUNN to which businesses can apply for up to 50 hours of assistance from a researcher to develop an idea or get feedback on a problem. This scheme may be particularly relevant for SMEs, which often lack this type of in-house expertise. In AI, where there is a shortage of personnel with advanced skills, schemes like these where resources are shared can help more companies gain access to the expertise they need to start their projects or pilots.
Digital21 is a strategy developed by and for businesses. The goal has been to make recommendations to the authorities on how businesses can better develop and benefit from competence, technology, research and development in order to succeed in digitalisation. An important message in the strategy is that Norway cannot be best at everything, and should therefore give priority to the technology areas where Norwegian business and industry can profit most. Digital21 highlights artificial intelligence, big data, the Internet of Things and autonomous systems as particularly important technologies for Norway in the time ahead.
Business clusters represent a policy instrument that can be suitable for promoting business development in AI. Innovation Norway, the Research Council of Norway, and Siva currently fund a cluster programme. The business community takes on the roles of leading and coordinating the business clusters, but research institutions and public agencies are often members.
Cluster for applied AI
In Halden, eSmart Systems, the Institute for Energy Technology (IFE), Østfold University College and the research and innovation company Smart Innovation Norway have taken the initiative to develop a new cluster, the Cluster for Applied AI.
The ambition is to create an optimal platform for developing Norwegian companies' international competitiveness and sustainable social development through applied artificial intelligence. The purpose of the cluster is to create new jobs and sustainable development based on rapid development and application of AI. Key focus areas will be technological development, commercialisation, ethics, security and accessibility. The cluster will facilitate the sharing of data, infrastructure and other technologies that its members would otherwise be unable to invest in.
Source: Smart Innovation Norway
There are also good examples of the business sector launching its own initiatives. Measures such as the AI Village in Trondheim, where several companies have joined forces to leverage synergies and create a stronger community, are good examples of how cooperation within the business sector can produce positive effects.
Review of the system of policy instruments
In 2018 the Government launched a comprehensive review of the system of business policy instruments.31 The purpose is for Norway to derive maximum value creation and economically sustainable jobs within sustainable parameters from the resources that are channelled through the system of policy instruments.
EU policy instruments for AI targeting SMEs: Digital Innovation Hubs
Digital Innovation Hubs (DIHs) is a policy instrument that was launched by the European Commission in 2016 to help small and medium-sized enterprises (SMEs) to digitalise faster. The first generation of DIHs was created under Horizon 2020. Four Norwegian DIHs have been set up, affiliated with GCE NODE in Agder, SINTEF, Oslo Cancer Cluster and Digital Norway, all of which are active enablers for SMEs.
In the transition to the new EU programmes for 2021–2027, a new generation of larger and more binding DIHs will be introduced. These will be one-stop shops, and will cover an ecosystem (national and international) that can offer expertise and competence and contribute to enabling SMEs in particular to become digital and leverage the potential of artificial intelligence. Funding for the projects will come from the Digital Europe Programme, among others.
The DIH programme and Norway's catapult scheme share some common features. The Government has established Norwegian Catapult, a scheme that gives Norwegian industry access to test facilities in order to enhance innovation and value creation. NOK 125 million was allocated to the scheme in the 2019 national budget. Siva has given priority to further developing the catapult scheme in 2019, in cooperation with Innovation Norway and the Research Council of Norway.
In Norway the clusters will be key participants in the DIHs that are set up. One important task for the system of policy instruments will be to further develop complementarity between the hubs, the cluster programme and the catapult scheme.
DigitalNorway is a non-profit organisation whose mission is to make it easier for Norwegian businesses to succeed with digitalisation. DigitalNorway cooperates with research and educational institutions, business clusters and several of Norway's best competence and innovation communities. Among other things, the organisation provides tools for companies that want to make the transition from idea to product or service, and networks for sharing knowledge and experiences.
The Conversion Engine service helps companies build their in-house expertise in digitalisation and advanced production. Together with Smart Innovation Norway, NCE iKuben and NCE Kongsberg Innovasjon, DigitalNorway leads the work on the Conversion Engine for digitalisation while NCE Raufoss has corresponding responsibility for manufacturing/-advanced production. The aim is to help small and medium-sized enterprises throughout the country succeed in their digital transformation. Innovation Norway supports the programme.
Intellectual property rights
Protecting intellectual property rights is important for ensuring that the AI market develops in the right way. Any uncertainty about ownership of the various elements that make up solutions based on AI (data, development framework, pre-trained algorithms, etc.), how they are licensed or how access to the solutions is paid for, will have negative impacts.
The Government wants Norwegian companies to make informed and competent decisions regarding protection, use and enforcement of their intellectual assets and rights, and to have a professional and conscious approach to the way they handle the rights of others. Norwegian companies should secure the increased market access and value creation which professional protection and use of copyright law can afford them. Protecting their ownership rights can be particularly decisive in connection with internationalisation.
The Ministry of Trade, Industry and Fisheries has begun mapping competence levels and needs in the area of intellectual property rights in Norwegian industry, and will assess whether the guidance offered on the system of policy instruments is adequate.
Public agencies can find it particularly difficult to know how to deal with rights when development of an AI-based solution is conducted through cooperation between the public sector and a private company: for example, how should a public agency handle ownership and user rights in a partnership where a commercial party develops and trains algorithms using the public agency's data? This type of issue will likely arise in the future, and perhaps especially so in connection with using health data for commercial purposes.
One important exception here, of course, is open data, which may freely be used by both commercial and public actors, pursuant to the Norwegian Licence for Open Government Data (NLOD) or under a Creative Commons licence.
Google DeepMind and the National Health Service (NHS) in the UK
In the UK the National Health Service (Moorfields Eye Hospital NHS Trust) has cooperated with Google DeepMind on developing an algorithm that can identify eye diseases based on images of the eye. The agreement makes DeepMind the owner of the final system but entitles Moorfields to the right to use it free of charge for a given number of years after it is approved for clinical use.
Source: Moorfields Eye Hospital NHS Trust
Norway can contribute to ensuring that artificial intelligence develops in the desired direction by exerting influence on international standards related to AI. Such influence is exerted by participating in international standardisation activities and by chairing working groups in areas that are important for us. At global level, AI activities are conducted in the standardisation organisations ISO and EIC and at European level in CEN and CENELEC. It is most often the large companies that take an active part in standardisation activities. The threshold for participating in this type of activity can be high for many SMEs for various reasons, such as travel costs or the time involved.
Standard Norway has created a mirror committee (SN/K 586 Kunstig Intelligens) to follow up international standardisation activities in AI. The committee is composed of representatives from research, business and industry, national authorities and various interest organisations.
The Government will
- work towards enabling the Digital Innovation Hubs to help SMEs get started on applying AI
- engage in dialogue with Digital Norway on how it can enable SMEs to take more advantage of the potential of AI
- promote the work on international standardisation activities related to AI, particularly with regard to stimulating SMEs to participate in these activities
- develop guidelines, including proposals for standard agreement clauses, on how public agencies should deal with ownership rights when cooperating with the business sector on developing artificial intelligence
4.2 AI-based innovation in the public sector
The Government believes there is vast potential for the public sector to rationalise and create better services through digitalisation. Artificial intelligence is one aspect of this. In the future the public sector will use artificial intelligence to deliver more targeted and user-adapted services, enhance the social benefit of its own activities, rationalise operations and work processes, and reduce risk.
Artificial intelligence in the public sector can contribute to:
- more relevant advice and services to citizens in different situations in life
- better decision-making support for public-sector employees
- rationalising processes and optimising resource utilisation
- improving the quality of processes and services by automatically detecting possible deviations
- predicting trends based on data from both agencies and their environments
- natural language processing for sorting and categorising, and for translating between different languages and language forms
Use of AI in the public sector is still in an early phase, however. A survey32 shows that many agencies are still in the planning or testing phase, where they try to find out what AI can be used for. Some agencies have begun testing through proofs of concepts. The survey shows that more than half of the agencies regards organisational culture, legal and regulatory issues, and data protection and security as the major challenges with respect to AI. Lack of competence is highlighted as another challenge.
For example, agencies can find it difficult to assess which areas are suitable for AI. One area that raises several dilemmas is the use of AI for control purposes. Such controls can involve, for example, identifying individuals who may be violating regulations (that is to say where an algorithm identifies a high probability for this). For applications like these, consideration must be given to rule of law and protection against self-incrimination for individuals subjected to regulatory checks. The risk and consequences of false positives – i.e., of someone being wrongly identified and of the undue hardship this would impose on them – must be part of a data protection impact assessment, which must be included when a solution is under evaluation.
Such assessments related to the use of AI in public administration can prove challenging, particularly if the agency lacks the relevant in-house expertise. Uncertainty over regulations for processing personal data, relevance to the Public Administration Act or assessments of when it is acceptable to use IA may make agencies unduly reticent. This may deprive the public sector of important opportunities to improve its services or rationalise its processes. The Government will therefore ask the Norwegian Digitalisation Agency to prepare guidance on this matter, in cooperation with the Norwegian Data Protection Authority.
Projects using AI in public administration
The Government wants public sector organisations to facilitate experimenting with artificial intelligence to gain knowledge about and experience in the technology. Trial projects or pilots in AI will provide valuable experience that can be used when evaluating large-scale projects and can enhance understanding of the technology at all levels in the organisation.
Artificial intelligence is one of many important tools for developing sustainable public administration in both central government and the municipal sector. Cooperation and exchange of best practice across organisations will contribute to better insights into AI, and experiences gained by the large agencies, which often have their own IT and analysis units, can help smaller organisations get started on AI projects.
The municipal sector is in a unique position when it comes to potential cooperation, because all municipalities have an obligation to provide the same services to their inhabitants, which means they hold datasets covering the same areas. The possibility to share best practice and to cooperate on procurements and training measures may therefore be particularly important for the municipalities. Cooperation on data, algorithms and competence in AI may also create possibilities to see interrelationships across sectors.
The Government will therefore ask the Norwegian Digitalisation Agency to facilitate cooperation in AI for example by establishing common use cases/user journeys and sharing of best practice.
Examples of projects using AI in the public sector
Several public agencies have conducted projects in which they have used artificial intelligence:
Artificial intelligence in residence verification
The Norwegian State Educational Loan Fund (Lånekassen) has conducted a project using artificial intelligence to select students for residence verification. In 2018 Lånekassen’s residential verification process covered 25,000 students, of which 15,000 were selected using artificial intelligence (machine learning) and 10,000 were selected randomly (control group). The results showed that the selection made using machine learning was twice as effective at identifying students who had not documented previously submitted residence information compared to the control group.
Automatic posting of invoices
Customers of The Norwegian Government Agency for Financial Management (DFØ) spend considerable time and resources on posting incoming invoices due to uncertainty about the correct posting. In some cases this can lead to invoices being paid after the due date and can create extra costs for the agency in the form of interest and charges. DFØ is currently testing solutions from two companies, one of which involves an accounting robot that uses AI to propose the correct posting. The model is trained using historical data before making a prediction of the posting based on a combination of historical data and information obtained from the actual invoice.
Sources: Norwegian State Educational Loan Fund and Norwegian Government Agency for Financial Management
Interaction between the public and business sectors
Norway's public sector procures goods and services worth more than NOK 500 billion annually. These procurements can be used to promote innovation and use of new technologies. A study conducted by Menon33 shows that most public procurement processes are conducted without any market dialogue and without encouraging suppliers to supply systems that are radically new and innovative. The companies report that young, innovative companies in particular find it difficult to know what they need to do to win contracts.
An innovation partnership is a procurement procedure that facilitates product and service development through cooperation between buyers and developers/suppliers. Innovation partnerships are used for procuring solutions that are not currently available in the market. Preliminary experiences show that startups and technology companies win assignments in innovation partnerships more easily than in other public tendering processes.
One of the measures in the Government's digital strategy34 is to establish a programme for increased interaction between the public sector and start-up companies, modelled on similar programmes in the United States and the United Kingdom. Such a programme would also benefit companies whose activities are based on artificial intelligence.
The Government will
- develop guidance on responsible use of artificial intelligence in public administration
- facilitate cooperation and exchange of experience and best practice for AI in both central and municipal administration
- establish a new programme for interaction between startups and the public sector
- present a white paper on innovation in the public sector