In an exclusive interview for CABAR.asia, the founder of the zypl.ai Academy of Artificial Intelligence Azizjon Azimi talks about the problems of AI development in Central Asia, the prospects for its application in various fields, as well as in public administration.
Azizjon Azimi at the age of 19 became the youngest graduate of New York University with two diplomas with honors in economics and international relations. MBA / MPP with dual degrees from Stanford Graduate School of Business and Harvard John F Kennedy School of Management. He was included in the “30 under 30” list of Forbes magazine and became the first representative of Tajikistan in the history of this list. He is currently the CEO of the zypl.ai Academy of Artificial Intelligence, valued by foreign investors at 6 million USD.
Nargiza Muratalieva: Thank you for your interesting session within the framework of the CABAR.asia school of analysts and for agreeing to give an interview. As it is known, you already have many achievements, you also founded the first center for the study of artificial intelligence in Central Asia, which is now called zypl.ai. In this regard, there is a broad question – to which extent the artificial intelligence (AI – Artificial Intelligence – ed.) is used and developed in the countries of Central Asia and what difficulties do you face when you promote any projects related to AI and Big Data?
Azizjon Azimi: I think you have asked a very good question, because first of all we must formulate what exactly is needed for artificial intelligence to develop. Well, first of all, it is open data, which has high accuracy and a specific task. That is, we must build an artificial intelligence system in order to solve some problems: these can be business, social, government, and so on.
The second is computer capability (power). Here, as the world moves to cloud technologies, it is very important that researchers, especially in developing countries such as Tajikistan and Kyrgyzstan, have access to cloud technologies that would allow them to conduct research at minimal cost.
Third, and of course, I would say the most important – these are talents, personnel, people who understand what artificial intelligence is, know how to work with it. Therefore, these three things together lead to the fact that AI develops.
If we assess the development of artificial intelligence in different countries, we can say the following. In Kazakhstan, universities such as Nazarbayev University and the Kazakh-British Technical University are already providing ready-made specialists. And in Tajikistan, we are just starting to work on this, but we are starting on a much larger scale. In addition, in Tajikistan, a national strategy for artificial intelligence is under consideration by the government – this is the first such document in the region. If it is accepted by the head of state, then we will open the artificial intelligence department in 5 universities by next year. This already speaks of the scale and will be our achievement. Other countries like Uzbekistan, Kyrgyzstan, Turkmenistan are still a little behind in this regard. However, we hope that Tajikistan, if we speak from our point of view, will become an example of a country where AI will lead to rapid growth. This is our goal.
As we know, the quality of public administration in the Central Asian region, which is still in its learning curve, is far from the best. And the question arises, how can AI be applied in public administration? And can AI improve the quality of public administration in Central Asian countries?
Azizjon Azimi: You asked the right question, because in our region, for example, digital initiatives and digitalization of databases are actively developing. For example, in Tajikistan, the tax committee has a huge database. The problem is that these digital databases are the first step in the form of infrastructure, but the second step is that our analytics is a weak point and has a critical capacity.
Therefore, I think that we should be very careful with the use of artificial intelligence in government agencies. Because in such an area there may be cases when AI will start to be used for the bad side of society. Why? Here is the example of China, where artificial intelligence is used to compile social scores of citizens. For example, if you start criticizing the Chinese Communist Party on Facebook, your score will go down, and this can cause you problems. AI in this case will look at you negatively when issuing loans, when buying a car, or something else. Therefore, I think that in the case of government agencies, a clear regulation of AI is needed. And in our AI strategy, if you ask about Tajikistan, this is exactly what we wrote that the goal of the state is to make AI open, accessible, but at the same time adhere to the principle of people’s confidentiality in the implementation of AI.
You raised a very good question, which was just the next I had. Where is the line between ethics and the use of AI in our countries? Will there be any steps on some legislative initiatives in order to resolve this issue?
Azizjon Azimi: Yes, you have raised a very important issue now. Because the main distinguishing feature of what we are doing with artificial intelligence in Tajikistan and with initiatives in other countries of the region is that we have a whole part of the strategy devoted to how to make artificial intelligence ethical and humane. There is no such thing even in conversations so far in our neighboring countries. This, I think, suggests that we have a different approach, a different vision. We really want the privacy of citizens’ personal data to be fully respected when using, developing, and implementing the artificial intelligence. That is, I will give a specific case for you to visualize.
For example, we are now, in the framework of the AI strategy, talking about the fact that there will be a state portal of open data, where researchers, companies and institutions will be able to use the anonymous data of clients, that is, citizens, to do their research or projects. But we will definitely introduce a function within which any citizen can enter this portal and refuse to use his data, even anonymously. And this ensures that people, as owners of the data, will have control and authorization over who uses such data.
Another case is medicine, which is a very important case. Because who is responsible for the medical decisions that artificial intelligence uses? This is a question that remains open to regulation.
Thank you, very interesting. And here again, returning to international relations, probably one of the questions that concern specialists – how and can AI be used at this stage to either predict conflicts or for the purpose of peacebuilding? Is there any research in this direction?
Azizjon Azimi: Yes, I think it depends on the intentions and will of the people who are implementing such systems. In an ideal case, artificial intelligence, for example, could be used to automate the NGO registration process. Yes, in order to say or to identify there any cases of fraud in the banking system or in the NGO system. That is, again, artificial intelligence, if we use it correctly, to which results will it bring? It will bring us towards high efficiency, low operating costs and high precision. Accordingly, we can use these principles in different areas: in government agencies, and also in the development of international relations. However, the main point is that the will and intentions should be humane and ethical.
It is also interesting that in the West, as you said today during the session of the CABAR.asia school has already revealed that AI and ML technologies were used during the election campaigns. Will this future come to our region of Central Asia? Or has it already arrived, and someone is already using AI in electoral technologies? And how to avoid negative influence? Because, for example, in Kyrgyzstan there were cases when bots were tried to be used during the election campaign, and in order to increase the ratings of some leaders. Can AI somehow influence bots? That is, let’s say, automatically delete them?
Azizjon Azimi: Indeed, bots can open fake twitter accounts and simply automatically repost words there for one candidate or spread fakes about others. AI systems can be used to identify exactly the fake accounts that are doing this. Therefore, within the framework of the electoral process, such systems can be developed. They are already used in different countries of the world. I think it’s a matter of time until all this will come to our region. But I think it is again very important to understand that artificial intelligence is an additive. The main point is that the infrastructure itself should be correct, the electoral system itself should be correct, so that it works democratically. And artificial intelligence is like the icing on the cake, which will lead to some kind of optimization or some kind of improvement, but only if there are the right institutions, which lead to the choice (election) of people.
What trends would you highlight in terms of the presence of AI in our region and how it will develop? For example, I already see a trend that there will be some legislative changes, that is, they are already about to happen, and something needs to be done about it. How are these issues regulated?
Azizjon Azimi: Yes, you asked a very good question, because when we thought about the introduction of artificial intelligence, we evaluated the criteria in which sector of the economy it will be better to implement it. For us, the main criterion was the availability of data and whether there is a digital database at all. Because of this, in Tajikistan, the real choice was between the financial sector and the telecom, and we chose fintech, because the regulator in this area is more inclined. Accordingly, if we look within the framework of the region, then in Tajikistan we have already issued 2 million US dollars of microloans. We are planning to disburse 15 million USD next year. According to our calculations, almost every tenth microloan will go through our system. This is already a pretty significant result at the global level, so fintech, I think, is the number one sector where we will see the introduction of AI.
In addition, we have other developments in fintech: this is the recognition of passport data so that loan officers do not enter it manually, but simply take photographs automatically and see the passport data in the system; it is also face recognition. We also have such developments, but it’s all aimed at fintech. In Kazakhstan, which is very interesting from the experience of communicating with colleagues, it can be seen here that AI is developing more in the field of MedTech, that is, medical technologies. These are the Cerebra companies and others that are there, diagnosing the disease using artificial intelligence. Therefore, healthcare and fintech are two areas where we will see a lot of artificial intelligence deployments. And of course, telecom, because telecom knows the client better than the bank and hospitals. They really see our every step there on the phone and therefore, telecom companies have such a volume of data, such a database about a person. After all, they will also be the locomotives of AI implementation.
Thank you very much. And the final question will concern young people. What would you advise, recommend to our young people who are starting out? Well, what are your future professional plans?
Azizjon Azimi: In just a few decades, maybe 10-20 years, most of the professions in the world today will simply disappear. This is a very painful process of transition to a global economy, which I think is comparable to what took place during the first industrial revolution in the UK, when a lot of manual work simply disappeared, and people began to retrain. But this retraining is generally good for the economy: more productivity, productivity means more welfare in the economy. But it is painful for some people, segments of the population that are lagging behind.
Accordingly, if our youth wants not to lag behind, but to be ahead, they must at least be literate in the sense of what artificial intelligence is, how does it work, how can it be applied? And the first step is always data analysis. Therefore, I would advise our youth to delve more into statistics, data analysis, analytics in general. If, of course, they have an interest in programming, then they should definitely go to artificial intelligence. As for my professional data, I quite openly discuss this, that my goal is to make Zypl.ai the first Tajik Unicorn (Unicorn is a startup company whose capitalization exceeded 1 billion USD in a short period – ed.). This is the dream of our team, especially since there are no other Unicorns in the region, except Kaspi. But Kaspi is also a startup bank that existed as a bank. At the regional level, we have the motivation to become the first Unicorn. Let’s see how all this will develop further, but the ultimate goal for me is public service. Therefore, I see my experience in the private sector, rather as a cumulative experience, which will then allow me to use it for the benefit of society. But until we become Unicorn, we will not think about public service.
Thank you for an interesting discussion!