What problems can machine learning and AI solve in Central Asia? Which countries or stakeholders can contribute to the advancement of machine learning and AI? What are the challenges and risks facing societies with the gradual introduction of AI and machine learning into our lives in political processes? The editors of CABAR.asia discussed these and other issues with Professor of the Higher School of Economics (Russia) Evgeny Sedashov.
Evgeny Sedashov received his Ph.D. in Political Science from the State University of New York at Binghamton in 2019. Currently holds the position of Associate Professor at the Department of Politics and Management of the National Research University Higher School of Economics (Moscow, Russia). Evgeny Sedashov’s research interests lie in electoral studies, conflict studies, coalition politics, quantitative political methodology, and Russian politics. Evgeny Sedashov’s articles have been published in the scientific journals – Journal of Conflict Resolution, Political Analysis, Political Science, SİYASAL: Journal of Political Sciences. Evgeny Sedashov is the owner of a large number of educational grants, scholarships, and awards, including Sakip Sabanci International Research Award (2022), US Fulbright Fellowship (2013), scholarship from the V. Potanin Charitable Foundation (2010-2012).
How do you assess the current situation and situation in Central Asia with the use of machine learning and artificial intelligence in the field of public administration and politics in general, especially in the process of making political decisions?
In my opinion, there is progress in this direction. There are centers of expertise (for example, Nazarbayev University) and private initiatives (for example, zypl.ai). At the same time, I would still note that further steps are needed in order for the services market in the field of artificial intelligence in Central Asia to develop effectively and dynamically. I would suggest the following as possible steps.
First, it is necessary to attract international experts to work in the region. It is especially important to involve scientists and teachers in the fields of programming, applied statistics and data sciences. The high level of teaching at universities and the opening of educational programs related to artificial intelligence and data science will provide the market with highly qualified specialists in several years.
Secondly, it is necessary to develop exchange programs, within which students and graduates can get education in the world’s leading research centers in artificial intelligence and data science. Even one year of study in such centers will allow you to reach a new level of professional competence. As the number of specialists grows, so will the number of management decisions that use artificial intelligence technologies and advanced data analysis methods.
Which countries, donors, international organizations, or stakeholders can make a certain contribution to the advancement of machine learning and AI in our region? What projects or centers specialize in this area?
In my opinion, it makes sense to develop cooperation with all leaders in the field of digital technologies. If we talk about specific countries, then the choice here is quite obvious: the United States, Russia, and China.
Russia has programs to attract foreign applicants both within the framework of Rossotrudnichestvo (Federal Agency for CIS Affairs, Compatriots Living Abroad and International Humanitarian Cooperation) and at the level of individual universities. Leading universities (National Research University Higher School of Economics, Moscow Institute of Physics and Technology, Skolkovo Institute of Science and Technology) have created world-class educational programs that train specialists in data science and artificial intelligence. It is also possible to attract specialists from Russia for the implementation of specific projects in the field of IT.
In the case of the United States, cooperation can be expanded through exchange programs such as the Fulbright program. Government programs in which the state pays students to study at top universities in exchange for a commitment to return to the country are another promising direction. After studying in the framework of such programs, graduates will be able to make a serious contribution to the development of IT industries in their countries. Finally, there is China. With regard to China, I would recommend similar steps related to the development of cooperation in the educational field and the involvement of specialists.
How to combine the skills and knowledge of the humanities (liberal arts) with a programming language and machine learning? Is it possible to nurture a generation of such specialists in Central Asia? Should young people be motivated for this?
In my opinion, this prospect is quite real. Modern methods of machine learning already allow solving problems that were previously solved exclusively by humans. In particular, significant progress has been made in methods to automatically determine the tone of the text and the topics that are presented in it.
Data science and machine learning will increasingly permeate every area of our lives. Accordingly, knowledge in these areas may soon become a matter of professional suitability of a specialist.
I am sure, education that combines training in data analysis methods and machine learning, on the one hand, and meaningful knowledge of social sciences, on the other hand, is the trend of the future. A common problem is that specialists either have good methodological training, but do not know how to find interesting questions for research and analytical reports, or have high-quality training in the subject area, but do not always understand how this or that question can be translated into the language of data analysis. It is specialists who combine the best of both worlds that will be in demand in the future as academic researchers and leaders of management teams.
What problems can machine learning and AI potentially help to solve in Central Asia?
It seems to me that there is quite a lot of potential here in various fields. First, medicine. There is some progress in this area, but it is certainly insufficient. Thus, Kazakhstan occupies the highest place in terms of the quality of medical services among the countries of Central Asia – 62 out of 167 countries represented in the rating. The quality of medical services, especially in small towns, leaves much to be desired. An artificial intelligence solution such as IBW Watson could be a good assistant to doctors and improve the quality of diagnostics, especially in situations where there is no access to accurate diagnostic instruments.
Another possible direction is crowdsourcing.
In a situation where the monitoring resources of the state are limited, feedback from citizens is extremely important, allowing to quickly identify problems and respond to the most important of them.
The system of automated collection and ranking of such appeals in order of importance could significantly improve the quality of public administration.
Obviously, such systems exist, but here the question is precisely in the quality of ranking and determining the importance of the problem according to predetermined parameters. Also relevant is the issue of involving more citizens in the operation of such systems.
What are the challenges and risks facing societies with the gradual introduction of AI and machine learning into our lives, into political processes?
Most of the discussion in this area can be described as a trade-off between the increase in the quality of life associated with technology and the decrease in the private space of users of these technologies. Almost any application in a smartphone collects data about users, naturally actualizing the question of the ethical use of this data.
Machine learning algorithms are already pretty accurate at predicting people’s preferences. However, if an algorithm can accurately predict user preferences, wouldn’t that lead to a planned economy and the unfreedom associated with it? The Soviet economy was inefficient, not least because of the difficulty in processing large amounts of information. Within the framework of a market economy, the process of transferring and processing information was faster, and the quality of information was higher, because the consumer was directly connected with the seller. However, what if modern computing algorithms can process information quickly enough to make central planning efficient?
There are other risks that have already directly affected people’s lives. In political science, there is a phenomenon of the so-called “echo chambers”. The essence of this phenomenon is that, until recently, the news algorithms of leading technology companies adjusted the issue to the preferences of users. Liberal users thus read more and more liberal media, while conservative users read more and more conservative media. In the US, this, according to some researchers, has led to an increase in polarization and intolerance to the opposing point of view. Companies have subsequently adapted the algorithms to provide more diversified information, but this has not yet led to a noticeable decrease in polarization. In many ways, the very format of communication in social networks implies the emergence of micro “echo chambers” in which people of similar beliefs exchange similar opinions, creating the effect of an information bubble.
In conclusion, I would like to note that, despite all the risks, I believe more in the creative potential of data sciences and artificial intelligence technologies. Any technology has risks, the only question is how to create a system of institutions that allows to effectively manage these risks.