In May, Minister of Economic Affairs Mika Lintilä established a steering group to draft a proposal for a Finnish AI action programme. The group is led by Pekka Ala-Pietilä, Chairman of the Board of Directors of Huhtamäki Group.
The aim is no less than to make Finland a leader in applying artificial intelligence i.e., turning AI and robotics into success factors for Finnish companies and society
There is no single definition of AI and the steering group has no intention of getting bogged down in theoretical debate, especially when the focus is on getting most out of AI and related technologies. The importance of AI can be referred to the impact of electricity.
– The steering group has got off to a fast start. The aim is to get things moving rather than draft some long, boring report, says Kalle Kantola VP, Research at VTT.He is on the secretariat of the group.
Kantola says that much AI-based work is already being done in Finland. Such work is now being expanded and speeded up. VTT, for example, has a great deal of AI expertise. AI has also become a priority in VTT's latest strategy update.
– Finland has a great deal of structured digital data, because that's the nature of our industrial sector. We have long been making products for which performance requirements are important. That is why data has been gathered; in addition, the service business has become important.
Large amounts of data are also processed in the public sector, such as in healthcare. Kantola mentions research on premature births as an area in which AI could turn Finland into a world leader.
– Support can be provided for doctors when data is combined with other content. AI applications are still in their infancy in global scale, and Finland could benefit greatly from it if it does the right things, adds Kantola.
The key issue is application
Without data, there can be no artificial intelligence or the related business. Data is gathered via sensors and telecommunications, and the latter is at a good level in Finland. AI algorithms alonedo not create impact - Expertise and industry know-how are needed in addition.. Kantola points out that VTT has the whole package.
– We have expertise in business, i.e., the bottom line, and technology. This means that we can be at the forefront of helping various stakeholders. VTT's AI strengths lie in application domains.
Kantola divides companies involved in AI into two groups: appliers and developers. For companies applying AI in their business activities, AI is the next phase in the digitalisation continuum. The second group includes companies that produce AI-based solutions, such as image or speech recognition.
– Finland has many highly specialised small businesses that are developing AI solutions for certain technologies. Small and agile companies focus on developing solutions for the 'appliers'.
Appliers tend to be medium or larger -sized companies such as Kone and Cargotec. There are only a few globally large players which lead the AI development such as Google and Amazon, which have huge amounts of data.
– The world has a few large players, as well as many small ones that are developing technology in niche sectors. There is a lack of middle-sized AI companies. The purpose of the AI programme is to propose measures for the application of AI in particular, says Kantola.
He states that nothing will happen unless companies come on board, but they have already become interested. In global terms, the big players seem to want to grow their own ecosystems and are seeking partners. This also presents Finland with an opportunity.
VTT's AI strengths lie in application. – We can be at the forefront of helping various stakeholders, says Kalle Kantola, VP, Research.
An 'amazing race' on the international scene
Minister of Economic Affairs Mika Lintilä explains that AI has risen quickly up the policy agenda and has suddenly gone mainstream in international politics.
– No one mentioned AI when the Government Programme was drawn up in Smolna a couple of years ago. And plenty of intelligent people were present – in addition to politicians, Lintilä jokes.
Part of his recent face-to-face the related cooperation. Lintilä points out that an international 'amazing race' is underway in AI.
The USA is in the lead based on all indicators. According to the Economist magazine, four times as many AI-related patent applications have been filed in the USA as in Europe, and over twice as many have been filed in China. In terms of AI firms and their funding, the USA is the clear number one, with the largest players being Americans.
China announced a 'state initiative' on AI in July. This is a centralised programme based on which the pace of the general development and application of AI should be among the world's best by 2020 and AI should be a key force of renewal in industry and the economy by 2025. China aims to become the world's leading AI innovation centre by 2030.
Russia barely appears in comparisons. Instead, Japan, South Korea and India are quickly catching up with Europe.
– It is important to have clear aims. The earlier we get started, the more we can use AI to improve our performance. To me, it is clear that the successful municipality, region and society of the future will be the one best able to apply AI, says Lintilä.
He compares the advent of AI to the invention of the internal combustion engine.
– Over the next five years, AI will change our lives more than any other invention, just as the internal combustion engine once did.
According to Lintilä, Finland is highly capable of financing AI projects, despite the fact that the programme budget does not include earmarked funding.
– Innovative projects involving AI will have a high profile in funding by Tekes - the Finnish Funding Agency for Innovation, for example.
AI should be mainstreamed
The Minister of Economic Affairs points out that a key success factor for the programme lies in mainstreaming AI all the way down to the proverbial Joe Average.
– AI teachers and appliers are now worth their weight in gold. I have suggested to adult education centres that they should take this very seriously.
AI literacy is being discussed in the same spirit as media literacy has been for decades. However, there seems to have been no perceivable development in the latter. AI literacy does not refer to teaching people about artificial intelligence, but to everyone understanding how AI might benefit their work, whatever their profession.
Because AI will affect the everyday lives of everyone, not just a small group, the aim is to involve people widely. Fear has given way to reason in public discussions of the topic, and enthusiasm is now the dominant emotion.
In the context of society in general, AI challenges lie in the areas of legislation, support for innovation and access rights to databases. Ala-Pietilä's steering group is pondering questions such as who owns the data.
– We have heard claims that everything, from gene banks up, should be handed over to Brussels, but this is an area where we need to be protectionists. I want to keep these issues in Finland's hands and in the Finnish language, says Lintilä.
The AI steering group will run until the end of the current electoral term, but the minister expects it to continue beyond then.
AI really got started in the 1990s
Discussions of artificial intelligence go back to the 1950s, but its development has stuttered a couple of times due to unfulfilled expectations. The roots of VTT's AI expertise reach far back into the 1980s. VTT had three or four working groups involved in AI back then and in the beginning of the 90s.
– The end of the 90s and beginning of the next decade saw the term fall out of use. People referred to machine learning, if anything, says Research Professor Heikki Ailisto.
He is in charge of the 'How to make digital service ecosystems and platforms into success factors for Finland' programme, which is part of the Government's analysis and research programme.
Higher computing power and database growth are behind the new phase. Computing power can even be outsourced from the cloud, as can ready-made AI programs and platforms. This provides even small businesses with access to global markets.
For example, VTT is involved in developing the analysis of wrist device measurement data; Finns are world class in this area. Where wrist-top computers once had to be switched on when starting exercise, sport is now automatically recognised.
– Recognition of typical sports such as running, cycling, Nordic walking or football is now at 97 percent. Categorisation accuracy has improved fivefold in a short time. A wrist device can even distinguish between traditional and electric cycling, says Ailisto.
In occupational health care, sickness can now be predicted and prevented on the basis of simple questions, where an AI program compares the answers with those of 120,000 other people.
In the United States, sports commentators have already been replaced by journalist robots into which game stats have been fed. VTT and the University of Helsinki have tested AI in journalism in the reporting of election results. An AI program can produce basic news reports by polling area, announcing the election winners in the area in question.
In the forge rolling of steel, quality problems have been spotted by using AI to combine measurements from dozens of data sources.
Heikki Ailisto views the AI discussion as a quest for balance, since the first thing that AI brings to mind is the Terminator films.
– The idea of a sudden leap, where everything has changed and machines are genuinely intelligent is wrong. As is the idea that all jobs will disappear. But neither is it true that nothing will happen.
Ailisto points out that AI will also bring major changes to society in the long run, but not as quickly as claimed. The changes will take 10 to 20 years.
– For example, autonomous vehicles are already real, but they will not replace bakery vans or ordinary drivers by 2022. Instead, trucks may be moving around fenced off opencast mines without drivers, or driverless trade fair buses may cruise at walking speed within fair areas, or tractors on fields.
– On the other hand, even just slightly higher-end cars have a lane departure warning system, cruise control and emergency braking, through which AI is 'infiltrating' ordinary passenger cars. However, drivers will still be needed at the end of the next decade. But all town buses may be driverless in 25 years' time, Ailisto speculates.