In May, the American organization CNA (Center For Naval Analyses) published a report «Artificial Intelligence and Autonomy in Russia», which provides a comprehensive overview of the current state of civilian and military artificial intelligence in Russia.

ICT.Moscow discussed the report and the specifics of the industry with a lead expert Igor Pivovarov, chief analyst at the MIPT Center for Artificial Intelligence, creator of analytical report «Almanac Artificial Intelligence», organizer of the OpenTalks.AI conference, researcher and entrepreneur. He shared his insight on how foreign actors view AI in Russia, talked about the problems of this industry and possible ways to overcome them, the impact of the pandemic on the demand for technologies and the entry of Russian AI into foreign markets. Full interview in Russian is available at ICT.Moscow.

WHAT THE CNA REPORT OVERLOOKED

The study more or less clearly and accurately describes the state of affairs in Russia. Among the essential factors, the authors point out insufficient funding. They also note that in Russia the state is a major actor in the field of AI. I disagree with this point in particular, but they are right that some state-owned companies play a big role. Globally, the market is a driving force for AI technologies, because it is a matter of competition between companies, as well as their survival and business profitability. In Russia, the market is much less competitive.

On the other hand, they did not write at all about Yandex, which in my opinion is the number one company in AI, a trendsetter and a global leader in certain technological areas. The authors conclude that the main dynamics and vectors of development in AI are set by large state-owned companies, primarily Sberbank. Yes, Sberbank is a key driver in AI regulation, but we also have a lot of private companies that are engaged in and apply these technologies. All that was largely overlooked.

AI EDUCATION IN RUSSIA

Another roadblock that the report touches upon is the brain drain. Russia holds high standards for education, but we don’t know how to keep specialists from leaving the country. I was involved in the development of the national AI strategy and If I were to decide where to spend all the money, I would spend everything on education. I realized that education is Russia’s key strategic advantage, but its potential hasn’t been achieved yet.

In Russia, we have a fundamentally different educational system as compared to those in the West. Ours was built this way because of the state-planned economy we had in the past, which distributed a specific number of physicists, engineers, nuclear scientists, etc., annually. It turns out that we are initially stronger in fundamental education, but we lose in terms of the speed of adjusting today’s educational programs to present-day needs

According to the new Global Skills Report 2021 made by the educational online platform Coursera, Russia ranks 9th in the world in terms of competencies in business, technology and data science. Russian specialists demonstrate exceptionally high results in the field of computer programming, working with databases and web development. Last year, Russia was the country with the highest level of technology skills.

RUSSIA ON THE AI WORLD MAP

I’m afraid there are two reasons why Russian AI startups aren’t considerably visible abroad. First of all, objectively, Russia isn’t up front in this field. The AI ​​market here in 2019 was estimated at 291.5 billion rubles ($3.8 billion), which is approximately 2% of the world market. There are around 500 companies on the Russian AI map, and about 250 of them are startups, compared to about 5,000 AI startups in the world.

There is one more thing. Historically, Russian science mostly dealt with fundamental things, which often don’t even have any applied value. One of the most cited Russians in AI is Dmitry Vetrov. His team is engaged in deep mathematics in the field of neural networks. They make Bayesian ensembles, and they are broadly cited in academia. The media won’t cover their accomplishments, because they are doing something that lies much deeper than practical application. In the future, based on their results, developers will make neural networks with practical applications. In this sense, I have high hopes for our businesses. Yandex can offer a lot of great things for the media to write about.

WHAT CHANGED AFTER THE PANDEMIC

We have not yet published the data on the growth of the AI market in 2020, but over the year the market has grown significantly, both in Russia and abroad, precisely because of COVID-19. The pandemic has pushed mainly online services, and AI was a key factor. It has had an impact on healthcare, diagnostics as machine learning helps to classify images and react to them faster, predict what will happen next and shape the situation. In this sense, AI turned out to be in great demand, although it seems to me that it still has difficulties finding its way into healthcare. The medical community is too cautious and has very tough regulators.

AI IN HEALTHCARE

There are three big problems with AI in healthcare in Russia. One of them is data, which is complex, unstructured, erroneous, and personal. Everything is complicated there. Another big problem is certification. It is not very clear how to certify something that does not have a clear order of action or an algorithm. In Covid times, everyone simply had to close their eyes on this in order to get the result. In peacetime, the problem takes years to be solved.

The third problem is specific to Russia and generally very complex, which most of our companies have not even reached yet, but will soon: the payment problem. In this country, as in the rest of the world, everything is paid for through healthcare insurance, or compulsory medical insurance, in particular. The introduction of a new product for healthcare payment is always a difficult thing. The rest of the world at least knows how to deal with it, but in Russia, no one really knows. In this regard, I would say that the market is, in fact, too dependent on state policies. Something needs to be done to enable private companies to develop, while the state should become a transparent, fast and reliable buyer of services.

GLOBAL THINKING

Many Russian companies understand the need to become international, but selling abroad is not easy. To do this, you need to make a product that is competitive internationally, and for this, you need a general understanding of how to make such products. Overall, we have cases, we know that it’s possible. For example, we have the Prisma app that became quite popular globally. So, Russians are able to make good world-class products.

There’s also a certain psychological factor, not only in Russia but all over the world. The Chinese look at China first, the Americans look at America. Only rare countries such as Israel understand that there is no one to sell to domestically. In Russia, people look at the Russian market primarily.

MAKING RUSSIAN STARTUPS VISIBLE

I do not see any particular difficulties for Russian companies to enter the international market. On the other hand, today’s world is highly politicized, and, conventionally, when Huawei sells its cloud all over the world, people view it politically. I know for a fact that some of the biggest, high-profile international companies use Russian solutions, but they don’t really want to be public about it, since, apparently, this carries certain political risks for them. It’s quite possible that there are much more of such cases than we know of.

Grigory Sapunov’s company Intento is a good example of an international company. It’s an American startup made by Russians, an aggregator of cloud APIs. It offers customers its API and internally mounts it on the API of different clouds. Intento shows exceptional growth in America, and its service is popular among big companies. So we can see positive cases and let’s hope there will be more of them every year.

I’m afraid there are two reasons why Russian AI startups aren’t considerably visible abroad. First of all, objectively, Russia isn’t up front in this field. The AI ​​market here in 2019 was estimated at 291.5 billion rubles ($3.8 billion), which is approximately 2% of the world market. There are around 500 companies on the Russian AI map, and about 250 of them are startups, compared to about 5,000 AI startups in the world.

There is one more thing. Historically, Russian science mostly dealt with fundamental things, which often don’t even have any applied value. One of the most cited Russians in AI is Dmitry Vetrov. His team is engaged in deep mathematics in the field of neural networks. They make Bayesian ensembles, and they are broadly cited in academia. The media won’t cover their accomplishments, because they are doing something that lies much deeper than practical application. In the future, based on their results, developers will make neural networks with practical applications. In this sense, I have high hopes for our businesses. Yandex can offer a lot of great things for the media to write about.