1. Artificial intelligence is advancing at full speed
Finland launched its Artificial Intelligence Programme nearly two years ago in May 2017, when the Minister of Economic Affairs Mika Lintilä announced that Finland aims to become a world leader in the application of artificial intelligence and in new ways of work. During the past two years, artificial intelligence has become one of the most intensely debated topics in Finland and in the rest of the world. In this section, we examine how artificial intelligence is defined and who are affected by it. We also review the attitudes towards artificial intelligence in different countries and in international organisations. The implementation of the Artificial Intelligence Programme is detailed in section 3 (Eleven key actions ushering Finland into the age of artificial intelligence).
Background to artificial intelligence
There has been a great deal of discussion about artificial intelligence over the past few years. More effective and lower-cost computing capacity, vast increases in the amount of data that can be used by AI, and more advanced AI algorithms have all led to more intensive use of artificial intelligence. In fact, we are experiencing a new spring of artificial intelligence. Artificial intelligence has been referred to as one of the most important technologies of this century. It has been one of the most talked topics at business forums and has been at the centre of public debate.
Artificial intelligence is not a single technology as it covers a number of different methods, technologies, applications and schools of research. Artificial intelligence and AI methods, technologies and applications are part of the wider phenomenon of digitalisation. In the Artificial Intelligence Programme, artificial intelligence is defined as follows:
Artificial intelligence means devices, software and systems that are able to learn and make decisions in almost the same manner as humans. Artificial intelligence allows machines, devices, software, systems and services to function in a sensible way according to the task and situation at hand.
In the project Artificial intelligence and its capability assessment1 carried out as part of the Government’s analysis, assessment and research activities, it was emphasised that artificial intelligence is connected with many different branches of science: philosophy, cognitive science, linguistics, neuroscience, mathematics, physics, engineering sciences and computer science. In addition to technological development, artificial intelligence also impacts individuals and all areas of society. In that respect, it is examined from the perspective of moral issues, ethics, values and policies as well as social sciences, law, economics, and business sciences. The use and application of AI technologies can also be examined in a broad range of different sectors, such as medicine, trade, industries, police work and warfare.
The branches of science connected with artificial intelligence, the sectors using it and the moral issues, ethics, values and policies steering its development are shown in Figure 1.
Thus, artificial intelligence is relevant to a broad range of different branches of science, sectors of society and policies. It is also relevant to us as citizens and employees. A great deal has been written about the topic in newspapers and magazines, books on AI have been published and training in AI issues has been provided. Events for companies and professionals working in different sectors are also held, such as the monthly AI Monday, which is jointly arranged by a number of AI actors.
Unlike many other technology-related topics, artificial intelligence has also aroused keen interest among the public at large and policy makers. In two successive years, leaders of Finland’s largest political parties have taken part in the panel discussion on artificial intelligence arranged by the Finnish Information Society Development Centre and held at the Information Society Academy. Ministers of Prime Minister Juha Sipilä’s Government have discussed artificial intelligence at numerous strategy sessions.
It is characteristic of our time that there is enthusiastic debate on certain new technologies and people are very optimistic about their potential. AI experts may be irritated by the vague concepts presented in the debate and the non-professional approach to the issue. It is, however, important that people with only limited knowledge of the technology can also take part in the discussion. A good example of this is the project of Omnia AI Lab in which 100 senior citizens are trained as AI mentors.
It is essential to make people more aware of AI and at the same time raise the level of the debate on artificial intelligence, its potential, the challenges ahead and the choices that must be made at societal level. Initiating an extensive discussion, providing information and making conclusions have also been the aims of the Artificial Intelligence Programme.
We have also noted with pleasure that artificial intelligence is already part of day- to-day activities in an increasing number of companies. According to the SME barometer carried out during 2018, about 8% of all SMEs are already using or testing artificial intelligence and another 24% are examining the matter. In strongly growth-oriented SMEs, the figures are twice as high.
Finland’s Artificial Intelligence Accelerator project lists 15 Finnish companies that generate most of their turnover from artificial intelligence. Artificial intelligence is also at the core of operations in many other companies and organisations and in this publication, eight of them are presented in case reports.2
CASE KONE: PREDIC TIVE MAINTENANCE
As world cities and population are growing, more advanced technology is required each year so that moving in cities would be easy and safe. Public transport and other traffic, lifts and escalators must function smoothly so that large masses of people can move from one place to another without problems.
The Finnish company KONE, established way back in 1910, is one of the world’s largest companies in the lift and escalator sector and it has more than 1.3 million pieces of equipment in its maintenance portfolio around the globe. KONE wants to make human mobility easier; it uses a broad range of different artificial intelligence applications to ensure a smooth and seamless flow of people. With artificial intelligence, it is possible to predict and identify potential defects in lifts already before they appear.
The company has developed KONE 24/7 Connected Services, an AI-based setup for predictive maintenance of lifts and escalators. In 24/7 Connected Services, the pieces of equipment in the company’s maintenance portfolio are connected to the IBM Watson IoT Platform in which artificial intelligence helps to predict defects before they occur.
The lifts and escalators send data and the sensors installed in them measure specific matters and supply data on real-world phenomena. Using the data, the artificial intelligence application continuously learns how to deal with new situations and this helps the company to become more proactive as it becomes more familiar with different types of lifts and their maintenance needs. The service is used in the maintenance of new and old lifts.
The AI-based service has allowed KONE to make its products safer and the maintenance processes more transparent. The customers now know in advance how their equipment might be affected by defects and how they should be prepared for them in their maintenance budgets. Artificial intelligence has also made lifts smarter: lifts using the technology are able to communicate their maintenance needs so that the mechanics know what will happen before anything happens. As a result, malfunctions have become less common and the repairs can be carried out more quickly.
Last year, a hospital using the system sent KONE 25% fewer service request than in 2017. In other words, the predictive maintenance service has substantially reduced the number of malfunctions.
KONE plans to link nearly all lifts and escalators in its maintenance portfolio to artificial intelligence. The company is continuously investing in new technologies in such areas as data collection and processing as well as artificial intelligence.
KONE believes that the correct use of data will play a major role in the future: actors who can best use their data in the design and supply of services and products will have an advantage over their competitors.
A company operating in the lift and escalator sector Domicile: Helsinki
- More than 57,000 employees
- Established: 1910
- Turnover in 2018: EUR 9.1 billion
KONE is one of the world’s largest manufacturers of lifts and escalators. The company, which was established in 1910, also provides maintenance services for its equipment, which account for almost half of its turnover. KONE has invested in technologies ensuring easier and more effective maintenance and it uses artificial intelligence is such areas as predictive maintenance.
CASE CARGOTEC: WORKING TO MAKE MARITIME FREIGHT TR ANSPORT EASIER
The world’s population is growing, cities are becoming larger and energy consumption is increasing. We already know this. However, amid all this, global logistics and container traffic as part of it play a major role.
Surprisingly enough, there is plenty of room for efficiency improvements in global container traffic and many of the processes are based on communications between humans. For example, if there are unexpected changes to the arrival times of container ships due to weather, the information may not be automatically relayed to the cargo handlers waiting at the port and they are thus unable to change their plans. Or it may not be relayed to the lorry drivers who are waiting for the containers so that they could transport it further.
Cargotec has tackled this problem with the help of AI-based solutions. In container traffic, it is essential to ensure seamless cooperation between shipping companies, logistics companies and port operators. A more transparent value chain and quicker reactions to changes would mean less confusion, less wastage and shorter waiting times. Reducing the emissions generated by the equipment is also essential.
Energy modelling and fuel guarantee are the models currently in use. Cargotec supplies mobile machinery for lifting and moving containers at ports and terminals. When designing the equipment, Cargotec uses technical solutions maximising energy-efficiency and minimising emissions because fuel costs are a major cost item for operators who are increasingly committed to environmentally friendly and low- emission operations.
Cargotec has developed a machine learning algorithm that can forecast fuel or electricity consumption in each customer’s operations with a high degree of accuracy. Each customer is asked a few essential questions allowing it to receive an accurate forecast of its own future consumption. The forecast helps the customer to acquire the equipment best suited to its needs and to accurately predict the costs arising from the operations. A concrete picture of the future emissions is also produced.
Telemetry data collected from a broad range of different equipment provides the basis for the machine learning algorithm. Customers use their equipment around the world in different conditions in a wide variety of different tasks. The algorithm learns to understand how different types of equipment operate and how operating conditions impact fuel consumption and emissions. The model is so accurate that Cargotec grants a fuel guarantee, in which it promises to provide customers with a financial compensation if the consumption exceeds the promised guarantee levels.
More AI-based technologies and solutions for different fields will be introduced in the future. The equipment will become more autonomous and security solutions are assuming an increasingly important role. Artificial intelligence is playing a key role in this process.
- Cargotec Domicile: Helsinki Personnel: 12,000
- Established: 2006
- Turnover in 2018: EUR 3.3 billion
Cargotec is an example of an industrial company that uses artificial intelligence. Global maritime traffic and especially container traffic are growth industries and their processes are largely based on communications between humans. In practice, only the introduction of artificial intelligence has made improvements in them possible. The fuel guarantee granted by Cargotec is an example of the use of artificial intelligence. In this arrangement, AI-based solutions provide accurate estimates of the amounts of fuel consumed.
Artificial intelligence is conquering the world
The opportunities offered by artificial intelligence have been globally recognised. More and more countries have incorporated artificial intelligence as a key factor in their competitiveness strategies or prepared a national strategy for artificial intelligence.
The CIFAR research institute, which took part in the preparation of the Canadian strategy for artificial intelligence, has examined in its report the state of EU’s AI strategy and the national AI strategies of 17 countries at the end of 2018. In addition to the countries reviewed in the report, many other countries have also launched artificial intelligence work. According to the report, artificial intelligence has become a key theme in national strategies more rapidly than any other of the recent technologies.
In half of the strategies, funding for the measures have also been proposed, while the rest were documents providing more general guidance for political decision- making in the future. The priority in national AI strategies is on developing industries and enhancing national competitiveness as well as on research and access to expertise. Ethics and data policy are present in many of the strategies, while at the same time, future of work and use of artificial intelligence in government and public services are less often mentioned. The strategies of China, India and Russia are examples of extensive strategies, while France, Germany and the United Kingdom are examples of major European countries with extensive AI strategies. Such countries as Estonia and the Czech Republic are also preparing AI strategies.
In April 2018, the European Commission presented its own AI initiative in which the aim is to increase private and public investments, prepare for societal and economic changes and provide an ethical and legal framework for the development and application of artificial intelligence. The initiative was made more detailed through a coordinated action plan jointly prepared by the Commission and the EU Member States in December 2018. The plan is extensive and it contains infrastructure investments, measures to improve data accessibility, investments in research and application, development of ethical principles and international cooperation. Funding for the AI activities comes from the Horizon Europe and Digital Europe programmes.
In their own national AI strategies, major powers emphasise the economic and societal impacts of artificial intelligence but also the international competitive setting and the strengthening of their own political and military position. In the United States, companies play a major role in the development of artificial intelligence, whereas in China, the efforts are led by the state and in the EU, the priority is on citizens’ fundamental rights and ethical issues. This has also been described as the ‘third way’ of developing artificial intelligence. However, in the AI initiative presented by President Trump in early 2019, the emphasis is also on the development of artificial intelligence so that societal trust can be ensured. For example, measures will be taken to ensure more effective opening of public sector data resources for the use of artificial intelligence in such ways that protection of privacy, civil liberties and protection of confidential information will be ensured in accordance with national values.
The ethical issues associated with the use of artificial intelligence and robotics are common features in many of the international AI initiatives (such as those of the UN, the OECD, the EU, Nordic Countries and the Baltic States). In many of the initiatives, existing regulation is used as a basis for the regulation of artificial intelligence. In the initiative of the United States, the federal government is urged to coordinate the development of technical standards and steer the development of administrative standards. Accountability and security issues concerning artificial intelligence have also been discussed at international forums and in national strategies. The European Commission is examining product safety and liability issues from the perspective of legislation.
CASE ULTIMATE. AI: BE T TER CUSTOMER SERVICE WITH ARTIFICIAL INTELLIGENCE
Customer service is one of the largest business sectors in the world and it employs millions of people. However, studies have shown that customer service is among the most stressful sectors of the world because the pace of work is hectic and customers have become more and more demanding.
In fact, many companies have developed AI-based solutions, such as chatbots, to ease the workload of their customer service personnel. However, chatbot solutions are not yet sufficiently advanced so that they could replace customer service staff or provide customer-friendly service. The founders of ultimate.ai noticed these challenges
in artificial intelligence and customer service in 2016 and decided to start solving customer service problems with their own application.
Ultimate.ai is making customer service more effective using deep-learning artificial intelligence. This is a tool that helps a company’s customer service personnel during chats or email discussions. The software understands the customer’s questions and provides ready-made alternatives for answers, which the customer service employees can process and use during the customer service situation.
The method is based on deep-learning algorithms, which means that it is completely independent of such factors as the language of the customer service data. Unlike ordinary statistical methods in which ordinary algorithms must be taught the language structure, a deep-learning algorithm learns the structure when it is provided with customer service data. In other words, the algorithm can be provided with data in any language because it can learn and process a new language quickly and accurately.
The company has collected more than one million euros in capital funding from German-based Holzbrinck Ventures and Finnish-based Maki Ventures. The customers of ultimate.ai include such large companies as Zalando, ADAC and Finnair.
Its aim is to become a smart layer in a CRM system. The company wants to create an artificial intelligence application that can understand system text data, automate work and provide companies with an overview of the problems affecting their customer service.
The company’s long-term vision is to teach machines how humans discuss (to encourage artificial intelligence to learn from humans and to produce human- like discussion). When we have reached the stage where artificial intelligence understands and takes part in such discussions, the potential for using AI is even larger.
- Customer-service AI-startup Domicile: Helsinki & Berlin Personnel: 20
- Established: 2016
- Turnover in 2018: EUR 0.26 million
Ultimate.ai is a startup providing solutions for more effective customer service. Its AI- based tool assists customer service personnel in chats and other similar situations. The software understands the customer’s questions and provides ready-made alternatives for answers, which the customer service employees can process and use. The tool has already helped a number of large companies, such as Zalando and Finnair, to streamline their customer-service processes.
CASE TEQMINE ANALYTICS: ARTIFICIAL INTELLIGENCE AS A TOOL ENHANCING INNOVATION
Patents are the world’s most important source of information on scientific and technological inventions. There are more than 110 million patents in the world and about three million new patents are added to this total every year. However, the information contained in the patents is difficult to interpret and access. Furthermore, it would be nearly impossible to review millions of patents manually.
Until now, the common practice in innovation projects has been that a patent expert has been asked to give an opinion on whether the new idea is practicable or might somebody else already have invented it. However, there is now a disruptive and a more agile solution, which is based on artificial intelligence.
Teqmine Analytics is a Helsinki-based AI startup, established in 2013, that wants to promote human inventiveness and help people to make better and more effective decisions on innovations. Teqmine Analytics is offering a solution that goes through millions of patents with the help of artificial intelligence. Researchers and investors can put questions to the AI application (such as “Is the idea already patented?” or “Has somebody else already invented it?”).
The basic idea is to help inventors to make better inventions but the aim is also to help investors and strategy directors to make better technology-related decisions.
The patent analysis performed by the application is based on machine learning that can automatically check the uniqueness of a new patent application against the patents contained in a database of about 25 million patents. The artificial intelligence application learns more by going through about 100,000 new patents each month.
It helps to reduce the need for extra meetings and speeds up inventors’ work. The innovation has allowed companies to reduce the use of outside experts, which in turn has led to quicker decision-making.
Teqmine Analytics already has customers around the globe, from China to Silicon Valley. In Finland, its AI application is used by such customers as Wärtsilä, the University of Turku, VTT and the Finnish Patent and Registration Office. The company’s turnover is increasing by leaps and bounds because the application is in huge demand.
In the long term, the aim is to build an AI application that can act as an equal partner in all research and development teams of the world. The requirement is that humans can trust artificial intelligence.
High-intelligence products can only be added to the internet, if the users feel that they are in full control of what they are doing in the internet. For this reason,
exceptionally strong privacy protection and information security arrangements are built into the application. The system works on the company’s own private servers. Unlike a number of free services, Teqmine Analytics does not collect information or sell it to third parties.
The company realises that there is a global need for new inventions: scientific and technological creativity is critical to ensuring the future of humankind and life of Earth. For this reason, the company wants to make the service available to all people in some form and not only to those who can afford to pay for it.
- Teqmine Analytics
- An AI startup specialising in patent information analysis Domicile: Helsinki
- Personnel: 3
- Established: 2013
- Turnover in 2018: EUR 0.25 million
Teqmine Analytics is a Finnish AI startup, which has developed an AI platform for analysing patent and technology information. The aim of the company is to help researchers and inventors to innovate more effectively. There are more than 110 million patents in the world and they are the key sources of information in the development of new science and technology. Teqmine Analytics helps researchers, inventors and investors to use AI to interpret and understand the information contained in the patents and to communicate on it easily and effectively.
1 Ailisto et (2018) Artificial intelligence and its capability assessment. Publications of the Government’s analysis, assessment and research activities 46/2018.
2 The cases are as follows: KONE, Cargotec, Ultimate.ai, Teqmine Analytics, K-Group, Fiva, Neuro Event Labs, Avaintec.