Prof. Galya Angelova: Europe should become a leader in safe and ethical AI
Human labour will not become completely extinct, but there will be great changesStela Marinova , Sofia
The realisation that someone is watching you on the internet is an alarming one but true nevertheless, says prof. Galya Angelova in an interview to Monitor.
Prof. Angelova, what prompted the drafting of a framework for a national Artificial Intelligence Development Strategy?
The European Commission’s (EC) activeness in preparing the Digital Europe 2021-2027 programme is the primary factor that caused us to pay attention to the topic. If I had to describe our framework document in simple terms, I would say that it can be viewed as a summary of the EC’s vision and the main regulatory documents published over the past year and a half regarding AI development in Europe. Appended is data on where the rest of the countries are as relates to drafting their own strategies and what areas they are focusing on. I would not call the document a “strategy” per se because it lacks an assessment of the situation in Bulgaria or a plan of action. The framework refers to the country’s obligations stemming from its EU membership and what the EC intends to do to promote the development and application of AI. The document was prompted by the fact that a French team of scientists devised an AI vision for the country last spring.
What are the potential benefits from this plan, what areas could it impact?
Above all, we tried to shed some light on the topic. The term AI encompasses systems that observe their surroundings, analyse what is happening, learn from the processed information and can, to a certain extent, autonomously alter their behaviour to adapt and perform their tasks better. As most of the time these systems operate with a great volume of data, the EC recommends that common European spaces for data exchange should be organised to allow the accumulation of public data and therefore the design and testing of new data processing and analysis algorithms. The existence of big data is a crucial characteristic of this era of technological development. Nowadays, humankind generates more data in a year than it used to in 100 years. Consequently, we have new goals and ambition as relates to using this data. For example, you have a large number of satellite photos coming in and to process the information gathered you need supercomputers, on the one hand, and fast algorithms, on the other, so that useful conclusions are drawn from those. Technological advancements also allow for observing the development of crops via satellite photos or aerial monitoring. Such tasks seemed impossible not that long ago, but today we have the necessary technology and ways are being explored to create new services building upon the data collected. This example is an illustration of why we are already starting to talk about data-based economy. In addition, the framework document helps explain the so-called Digital Europe vision. The EC’s communications lay out several focus areas when it comes to AI development. The first one is public data and the common European data spaces. All Member States have to set up clouds where public organisations can store data generated over the course of their work, including results from public-funded scientific research. The development of healthcare is also tied to data.
The EC is endeavouring to create large storages of anonymised medical data and so enable analysis that should reveal more about diseases, treatments, drugs or how a given illness progresses in general, thus facilitating medical breakthroughs. In this area, Bulgaria has some catching up to do not only with regard to data collection but building electronic medical records of patients and systematically gathering genetic information about citizens, which is a part of modern medical study. The question of data is one of the major points of discussion when it comes to AI in Europe these days. Education and science are another highlighted area. Member States plan to increase the number of university departments dealing with AI (Germany intends to add 100 professorships and attract experienced experts from around the world), as well as integrate classes in other disciplines so that young people have an idea how to process data in their own specialities, thus making AI a part of interdisciplinary programmes. The establishment of AI excellence centres across Europe is also envisioned, with the EC committed to help foster strong connections between these centres through projects under EU-funded scientific research programmes. In this way, not only AI technologies will be developed, but the implementation of applications in various economic sectors will be facilitated. The third area on which Europe is focusing is the ethical and legal aspects and the anticipated change in the labour market. The accelerated automation process is resulting in the loss of jobs as workers are replaced by machinery or software. This affects not only manual labour but services as well.
Could you give an example?
Voice assistants ran by AI-powered software are increasingly being introduced in call centres – you talk to them and they answer. That naturally raises the question what will happen to the people who do this for a living. Even though human labour will not be completely replaced in occupations that involve communication and empathy, significant change is on the horizon. And so we get to the problem of job retraining, which will be a major factor in the coming 10-20 years. The rapid development of technologies will also require retraining. One thing that is being discussed is that all countries should have vocational training plans that would allow citizens to respond to the changing labour market. In Denmark’s strategy, for example, one of the main goals is for every citizen of the country to develop basic digital skills. Even there the digital literacy rate of the population is below 100%, but it is clearly above that in Bulgaria. This takes us to the need for lifelong-learning programmes to be organised so that the process of retraining and acquiring digital skills can be made much smoother. It is believed that older people will find it harder to adapt because unlike the young generation, who are starting out at a school age, a person trying to understand the digital world in their 40s or 50s will be more challenged. In connection to the selection of priority topics, I would like to add that many European countries identify agriculture as one of those when it comes to introducing AI (in addition to healthcare and transportation). Self-driving farming machinery that can perform certain tasks with the help of drones is becoming available. Animal farms with a high degree of automation are cropping up.
What should be done in order for such a strategy to be drafted in Bulgaria?
I think that all strategies should start with a holistic analysis of the state of the involved sectors. There are numerous documents already prepared, like the concept for digital transformation of the Bulgarian industry (Industry 4.0), which shows that internet connection is excellent in the cities and not so much in the more remote settlements and that more and more people use digital services, although the percentage is lower among older people. We also have a draft Strategy for Digitalisation of Agriculture and Rural Areas. These preliminary studies should be considered in conjunction and several priority areas should be selected for Bulgaria, as it is impossible to fund everything at the same time. This has been the approach taken by the other Member States. I did not specifically mention education and science as priority areas because their development is simply necessary in order to have well-qualified employees.
What measures should public institutions take?
Coordinated measures across various sectors should be planned and geared towards one goal – to foster implementation of new technologies, which will improve the economy’s competitiveness and allow Bulgaria to attract high-tech investors. One of these measures is refining the national legislation. For example, the introduction of autonomous robots in real life presupposes differentiating between responsibility shouldered by the manufacturer and by the consumer who buys the device and uses it. Until sound legislation is put in place and consumers come to trust that they will be protected from malicious use and that they understand at least in theory how the AI systems make decisions, the new technologies will not be embraced. The public should be engaged in dialogue and given examples of the benefits from AI technologies. I believe that Bulgaria faces some serious and specific problems like the large number of young people who neither study nor work. The state should consider their further education and qualification and what the labour market will look like in 10 years.
How does Europe stack up against the world?
The European documents stress the strategic role of AI as a 21st-century technology that can improve Europe’s competitiveness. Manufacturing costs go down as a result of replacing the more expensive human labour with machines, which is why Europe should do its best to transition to modern technologies as soon as possible. China provides serious competition in that regard, as does the US. Those two countries invest staggering amounts in introducing AI into practice. Two months ago Taiwan announced that it will be training 10,000 AI experts per year in an effort to entice large Silicon Valley companies to open subdivisions in Taiwan and make their products in Asia. We do not seem to be in the mix. The EU is striving to become a world leader in ethical and safe AI, which, in addition to improved competitiveness of the economy, entails better quality of life for European citizens. The strategies and goals of different countries vary in approach, with the European ones strongly geared towards citizens.
Is there a danger of Bulgaria becoming isolated then?
No, we are a part of the modern world. Albeit slowly, we are meeting the requirements and guidelines for the development of digital Europe, and young people definitely feel very comfortable in this environment. In fact, we use AI on a daily basis even now, as adaptive systems are already here. For example, if you buy books or plane tickets online, the internet browser starts showing you ads of other books or hotels in your destination. In other words, the browser tracks what we do and adapts to our preferences. The realisation that someone is watching you on the internet is an alarming one but true nevertheless. There is no culture of talking about data in Bulgaria and we do not have large volumes of systematically provided public data. We are yet to pay attention to the fact that products or services can be built on this data, profits can be made, which means job creation. This is one of the things we should talk about in order to grasp the importance of data.
Prof. Galya Angelova is an alumnus of the Faculty of Mathematics and Mechanics with Sofia University “St Kliment Ohridski”. She earned her PhD at the Institute of Automation and Computing Machinery with the Hungarian Academy of Sciences (MTA SZTAKI). Since 2012, she has been a professor in informatics at the Institute of Information and Communication Technologies with the Bulgarian Academy of Sciences. In 2015 she received the Pythagoras Award of the Ministry of Education and Science in the Successful Manager of International Projects category for her work on coordinating the AComIn (Advanced Computing for Innovation) project.