90 billion rubles for the development of artificial intelligence

    On May 30 this year, a meeting on the development of technologies in the field of artificial intelligence was held on the territory of "School 21" of Sberbank. The meeting can be considered a bit of a milestone - firstly, it was led by Russian President V.V. Putin, and presidents, general directors and deputy general directors of state corporations and large commercial companies participated. Secondly, neither more nor less was discussed, and the National Strategy for the Development of Artificial Intelligence Technologies , prepared by Sberbank, about which G.O. Gref.

    The meeting seemed interesting, albeit a long one, for almost an hour and a half, so I offer a kind of digest of the main statements and opinions of the participants. Quotes were chosen the most key, it seems to me, on the topic, so as not to dig into the details. The numbers in front of the names of the speakers indicate the time code for the video, links to the video are at the end of the article.


    05:10 Vladimir Vladimirovich Putin, President of Russia

    [...] Today I propose to discuss specific steps that will form the basis of our national Strategy for the development of artificial intelligence technology.

    [...] This is truly one of the key areas of technological development, which determines and will determine the future of the whole world. AI mechanisms provide real-time quick adoption of optimal decisions based on the analysis of gigantic volumes of information, the so-called big data, which gives a huge advantage in quality and performance.

    [...] the struggle for technological leadership, especially in the field of AI, and you all know this very well, dear colleagues, has already become a field of global competition

    [...] if someone can ensure a monopoly in the field of AI - well, the consequences we all understand - he will become the ruler of the world

    It is no accident that many developed countries of the world have already adopted their action plans for the development of such technologies. And we, of course, must ensure technological sovereignty in the field of artificial intelligence. [...] We need precisely universal solutions, the use of which gives the maximum effect, and in any industry.

    To solve such an ambitious project in the field of artificial intelligence technologies, we objectively have good starting conditions and serious competitive advantages. [...]

    13:04 German Oskarovich Gref, Sberbank

    [...] during this period of time we have developed not only the document itself called “Strategy”, but it is very important that we also managed to create a document called “road map” to this document. In general, today we have two draft documents that can be considered an integral document, which is subject to approval.

    [...] In 2017, five countries adopted the National Strategy for the Development of AI, and during 2018-2019, 30 countries already. If this document is approved in the near future, we will be the 31st country to build a road map and declare a priority in their activities.

    [We include in the concept of “Artificial Intelligence]]
    • Computer vision
    • Natural Language Processing
    • Speech recognition and synthesis
    • Recommender systems and intelligent decision support systems
    • Advanced AI methods and technologies (primarily AML - automated machine learning technologies)

    [...] In the process of developing the Strategy, we identified and analyzed six driving factors for the development of artificial intelligence.
    • algorithms and mathematical methods;
    • software;
    • data, work with data, regulation and use of data;
    • Hardware;
    • everything related to education and personnel;
    • regulation

    Each of these six factors is critical. The absence of one of them creates critical risks for the system as a whole.

    In each of the areas we have set goals.
    • Algorithms and mathematical methods - by the age of 24, enter the top 10 countries by the number of conference participants and by the year 30 enter the top 10 countries by the average citation level
    • Development of software and technological solutions - to develop solutions that can provide superiority over a person in special tasks, and by the year 30 we must ensure superiority in a wide range of tasks
    • Storage and collection of data processing - create an online platform with anonymized state data and company data, to which AI system development companies will have access
    • Specialized hardware - the creation of their own architectural capacities that can create the architecture of the corresponding chipsets, and, accordingly, a specialized production site that can produce them.

    • Training - we want to enter the top 10 countries in educational programs in the field of artificial intelligence by 2024. And by 2030, eliminate the shortage of specialists in the field of artificial intelligence.
    • Creating the correct normative regulation in the field of AI is important here to go between two extremes: not to leave this sphere unsettled, on the other hand, to nevertheless create opportunities for it to maintain its development dynamics.

    We ask you to approve the National Artificial Intelligence Development Strategy with your decision, create an appropriate coordination body, and instruct the Government of the Russian Federation to approve the “road map” for the development of artificial intelligence.

    31:54 Maxim Alekseevich Akimov, Deputy Chairman of the Government of the Russian Federation

    [...] What we are definitely not going to do is another bureaucratic construction. We think that organizational tools have been created as part of the national Digital Economy program. And having concentrated the resources that are currently available for this program within the framework of a separate federal project “Artificial Intelligence”, we will cope with the strategic challenges that German Oskarovich spoke about.

    [...] What activities could include this plan with total funding, in our estimation, over a six-year horizon of up to 90 billion rubles? ... It is necessary to create favorable conditions for the development and replication of technologies, to subsidize pilot implementations, because this is a risk that private companies can and should share with public entities. Therefore, we will provide the leading companies with resources for prototyping the use of artificial intelligence, creating the basis for future implementations.

    [...] The public procurement system today is infinitely far from not only purchasing technological solutions by the public sector, but also for streaming development by means of fast tracks based on modern principles of software in general. And in this regard, Vladimir Vladimirovich, I would ask your instructions to the Ministry of Finance, the Ministry of Economic Development, and the Government to deal with this issue. Special regulation of this area is necessary.

    [...] Together with the deputies of the Federal Assembly, we have developed amendments to the law on personal data specifying mechanisms of depersonalization.

    [...] update industry standards, including in the field of security [...] predictive analytics of the operation of complex equipment [...] will create opportunities for a transition to a risk-based approach in control and supervision activities.

    [...] In the future, all specialties related to the adoption of significant decisions will require competencies in the field of artificial intelligence. And much work remains to be done on educational standards.

    [...] Mass and deep training of civil servants is also necessary. [...] we will promptly form an additional federal project, prioritizing expenditures. And we plan to do this no later than October 2019.

    [...] We propose to determine the Ministry of Digital Development, Communications and Mass Communications as the responsible federal executive body.

    [...] On the second question:
    [...] It is large Russian business that should and can be the main participant in the development of high-tech industries in the Russian Federation
    [...] In coordination with companies, the following distribution of topics is proposed.
    • The leading company in the direction of AI will be Sberbank
    • 5th generation mobile communications technology - Rostelecom, Rostec
    • Quantum Sensors - Rostec
    • Distributed Registry Technology - Rostec
    • Broadband for the Internet of Things - Rostec
    • Quantum Computing - Rosatom
    • New materials - Rosatom
    • Quantum Communications - Russian Railways

    During this year we will receive and proceed with the implementation of detailed roadmaps.

    42:20 Sergey Semenovich Sobyanin, mayor of Moscow

    47:30 Kirill Alexandrovich Dmitriev, RDIF

    [...] We analyzed 100 leading companies in the field of AI, selected the 20 most promising, and already approved financing in 6.

    [...] Eight brief sentences:
    First. There are two data management models, the Chinese, where the state provides greater access and control over data. Another European, where access to data is more limited. It is the Chinese model that will allow us to move forward faster.

    The second one. Bringing Russian companies to the level of world leaders. Because, if our companies work only on the Russian market, then they do not have enough advantages to compete in world spaces. And we want to bring our companies to world markets.

    The third. It is important that our companies implement AI, and we believe that each of the large state-owned companies should have its own strategy for the implementation of AI

    Fourth. It is possible to build consortia with China, the Middle East, where it is useful to have large-scale markets

    . Fifth. Create, together with Moscow State University, the

    Sixth Artificial Intelligence Development Center . Datacenters

    Seventh. There is really a lot of data both in Moscow and including the Federal Tax Service, which have significant potential for using the economy, and these data can be used

    We propose supporting the inclusion of Sberbank in the consortium, and we recognize that Sberbank was one of the first to focus on AI, and RDIF and Gazpromneft, which have very good achievements in this field so that they can jointly develop and promote this.

    51:08 Alexander Valeryevich Dyukov, Gazpromneft

    [...] A fundamental factor in the development of AI is demand. Demand needs to be formed, it needs to be consolidated so that the AI ​​development strategy makes sense. Now conscious demand exists in such areas as banking, media, retail, telecoms. But demand in these segments is still limited

    [...] The fuel and energy complex has a number of advantages over other industries necessary to become a leader in the development of AI. The fuel and energy complex is able to form a solvent demand for the technology of introducing AI technology ... tasks solved by the companies of the fuel and energy complex can be replicated and scaled in other industries.

    [...] We are ready to work together with other companies and organizations to organize the strategy, and we are ready to take on the role of one of the leaders in the development of AI for the industry segment

    55:56 Sergey Viktorovich Chemezov, Rostec

    [...] the development of specialized hardware - in this direction we have already begun to work - created a private-state partnership with AFK Sistema, combining all assets [...]

    As for the whole AI, yes, I want to emphasize that this enterprise is not closed, and we are ready to accept, include in this enterprise everyone who needs to develop this sphere of activity, who have any achievements, so please. I know that we have Angstrom-T, in the future I think this company can also enter our community. [...]

    1:01:06 Yuri Ivanovich Borisov, Deputy Chairman of the Government of the Russian Federation

    [...] I would like to see in the Strategy, of course, along with important goals, such as entering the top 10 countries in terms of the number of articles and participation in conferences, as well as in the average citation rate, to see more goals related to the share market, global and domestic. It seems to me that the main objective of this strategy is the dynamic implementation, I emphasize, of domestic AI solutions - this is just a complex of algorithmic, software and hardware solutions in all areas of the economy with the goal of dominating these products on the domestic market, and the forecast shows that this is a huge market , and positioning on the outside.

    [...] Of course, the activities to be planned in the Strategy will be aimed at creating these products, and this is good, I hope so. But it seems to me that the main objective is the promotion of these products on the market, and I completely and completely agree with Alexander Valerievich that the main objective of the Strategy is the formation of demand.
    Given the fact that these products are expensive, you may have to think about the state order for the introduction of these products in specialized centers [...]

    1:03:43 V.V. Putin

    I fully subscribe, I agree, because the goals that we set for ourselves should be, including if they are not very tangible at first glance, but we must strive to make it possible to measure the result of our work, this is true.

    1:04:03 Arkady Yurievich Volozh, Yandex

    [...] I would like that, within the framework of the strategy, perhaps, a special state program should be developed that stimulates not only materially, but maybe immaterially people returning to work here. Other countries have such programs, we also need to do this.

    [...] And the second aspect is the creation of conditions for testing. [...] you need to train cars with real data, in real conditions, for example, drones, unmanned vehicles that go along roads. They should go not so much at the training grounds as they should go onto the roads of street use, and here it is very important that we do not regulate this. [...]

    Yandex needs to bring a hundred cars to the streets this year. If we take the procedure that is now, we just need four years to certify these machines. I want it to be reflected in this program as well - tests in a real environment. Because we are talking about the fact that we will import or export these technologies in the end.

    1:08:08 Dmitry Nikolaevich Peskov, Special Representative of the President of the Russian Federation on Digital and Technological Development

    [...] Today we rely on the data that has already been created by banks, telecoms, and the state. But there are much larger data arrays, and we see that world development today is turning in the direction of such completely unexpected types of data - these are oceans, forests, humans, biomes, microbiomes. We see a huge number of startups that today compete for traditional markets on the logic of the link between biology and artificial intelligence. Brand new types of products are emerging.

    [...] This joint, it seems to me, should be finalized before the formation of the DataSet, not only from banks and telecoms, but also in the setting of tasks in scientific research, in industries, in the forest industry, in many other places from the point of view of formation new types of DataSet.

    [...] The second is topics related to personnel and regulation. The gap in personnel requirements is now so catastrophic that we could not find a single scenario in which we could close the gap even before 2030. In this sense, if we continue to rest on the idea that we can modernize the existing system of educational standards, we will never get the result.

    [...] We need a separate regulation area on the topics of artificial intelligence of big data and other cross-cutting technologies, which would allow us to normatively launch training processes for two to three months, and not for three, four, five, six years. Once again: there is not a single scenario.

    What is the problem here? The problem here is that data is really feed. Around the data there are many new specialties. [...] this is a deepening and division of labor, it requires separate regulation. I would ask separately in the instructions we still need to form such a contour; it is not yet complete.

    Last: of course, the entire education system must be changed. Today, we are taking such a step with our colleagues from the Ministry of Science and Higher Education, we are starting the implementation of the artificial intelligence educational platform right away in one hundred regional universities and we hope that it will work already in July this year. I also invite everyone to take part in its launch.

    1:12:30 Mikhail Eduardovich Oseevsky, Rostelecom

    1:13:25 Boris Oleg Dobrodeev, Mail.Ru Group

    [...] We, the company, would be very interested to participate in its implementation. For us, artificial intelligence is no longer the future, it is the present. More than 100 million people today use our services based on artificial intelligence. And, of course, it would be very interesting for us to apply this expertise in the real economy.

    [...] Every day and inside, we create a large number of services and look at hundreds of startups, and we are convinced that the main issue of this market is just the issue of the sales market. Because now there are not dozens, even hundreds of companies that have good technologies, but they all rest on the sales market, on small revenue, which today does not pay for high-tech startups and technologies. Therefore, it seems to me that the most important goal for us is precisely the stimulation of this demand and sales markets.

    1:14:39 Ivan Mikhailovich Kamensky, Rosatom

    [...] I would love to ask you today to support what German Oskarovich is today

    [...] But I wanted to support Yuri Ivanovich that the most important task is to create a market, an internal market and an external market for the implementation of these technology.

    1:15:45 Andrey Removich Belousov, Assistant to the President of Russia

    [...] What Ivan Mikhailovich said and what Yuri Ivanovich said about the market [...] But the market for artificial intelligence products, firstly, does not exist as such, and secondly, it is absolutely not the main thing . The main thing is the change in traditional markets, which carries artificial intelligence, and we need to measure not what share we will occupy in this market of artificial intelligence, but how much we can win back through our national decisions in the industrial market, in the market of trade services, in the market of logistics services - that’s to be measured.

    Operationally, of course, these are not meters, we can never measure it. We mean this, realizing that this is the main effect, but to draw some numbers for the 24th or 30th year, you yourself understand that this will be somewhere on the verge of speculation.

    [...] about the fact that there are sales problems and so on. They will continue to arise. [...] We must understand with you that the introduction of artificial intelligence fundamentally changes the decision-making processes in companies. It is impossible to implement artificial intelligence without changing the control system. It just won't work.

    [...] the most important mistake that we can fall into is to begin to form some new control systems, a new design for controlling this process in the strategy, [...] therefore the solution scheme that was voiced by Maxim Alekseevich is now proposed here : the existing decision-making system that has developed within the framework of the Digital Economy program, we are making changes there, but changes that will affect the content in many respects, but, thank God, will affect the bureaucratic form to a minimum in order to to save time. [...]

    1:20:15 V.V. Putin

    [...] Of course, when introducing the achievements of artificial intelligence, of course, we need to provide our companies with a natural capture of the market: both their own market and global markets.
    At the same time, there are such areas of activity, say, the use of artificial intelligence elements in medicine [...] Today we talked about how the work in this area is improved with the use of artificial intelligence - by 30-40 percent. You can use it, but you can not apply. You see, you can not apply. So far we have lived like this, and nothing seems to be like. And in order to be applied, appropriate decisions of ministries and departments are needed. Still, we need to promote these products.

    [...] And here a lot also depends on ministries, including the Ministry of Industry. Either we set certain requirements for the application of known achievements, or not, and everything will roll like that. [...]

    [...] You’ll probably have a good strategy. I just want to draw attention to the fact that, in principle, we can write strategies, even the most complex ones. This is a complex strategy, we need a step-by-step plan for implementing these strategies, in this case, a step-by-step plan for implementing the development of artificial intelligence. This will definitely need to be done. Maxim Alekseevich has already talked about this, but it is necessary so that it is still understandable and clear how this will move.


    In conclusion, I want to talk about my own impressions of this meeting.

    The first thing that catches your eye is the use of the term “artificial intelligence” to describe a set of technologies where AI does not smell, the more powerful AI . Here, “artificial intelligence” is nothing more than a beautiful wrapper, under which they sell mainly machine learning. But in reality, at present, there is no certainty that humanity can create a strong AI in any clearly predictable period. And, most sadly, the future national strategy is not going to solve this problem - a strong AI simply does not fall into its sphere of interests. And with the development of machine learning, the big question is also whether there are, or at least possible, more effective methods than known.

    The second, the meeting, identified two different goals that their supporters suggest should be a decision to develop AI technologies. In the Strategy prepared by Sberbank, Russia (and Sberbank) is expected to gain a reputation as a world technological leader in the field of AI, and the other part of the participants has an opinion voiced by Yu.I. Borisov, that the main goal is to develop and implement our own AI technologies in our economy, thereby raising both technology and the economy to a new level. The problem, however, is that the achievement of Sberbank’s goals can still be measured somehow, but the percentage of technology development and economic growth based on them cannot.

    Third, the opinion of Yandex and Mail.Ru, the companies that are really engaged in these technologies, their implementation in practice, and the development of specialists, was interesting. It seems that the money allocated for development will pass by them.

    The full video can be found on the Sberbank TV channel on Youtube and on the website of the President of Russia . The second link also contains a complete transcript.

    Also popular now: