RIAC FORECAST 2019–2024
Global Issues
RIAC FORECAST 2019–2024
Prospects for the Development of the Space Industry Until 2024
Alexey Fenenko
Doctor of Political Science, Associate Professor at the Faculty of World Politics of Moscow State University, RIAC Expert
About a decade ago, the term "new space race" was popular among space experts. Coined by President of the United States George W. Bush in 2004, this denoted the start of a new round of competition among great world powers for the exploration of near and deep space. The United States put forward ambitious projects for the unmanned exploration of the Solar System (including the Sun) and manned missions to the Moon and Mars. Russia, for its part, stepped up work on its constellation of GLONASS global positioning satellites and was planning its own deep-space research missions that would cost less than the NASA projects. China successfully carried out its first manned orbital flight, continued to expand its satellite constellation at a rapid pace, and was developing its own programs for the exploration of the Moon and Mars. The European Space Agency (ESA) created ATV unmanned cargo vehicles and began mapping the Moon, Mars, Venus and the satellites of Jupiter. Second-tier spacefaring nations such as Japan, India, Brazil, South Korea, New Zealand and even Iran all attempted, with varying success, to develop their own suborbital launch vehicles. The authors of a number of publications compared this period to the U.S.–Soviet space race of the 1960s. [1]

Indeed, during several years, the existing spacefaring nations did attempt to recreate the spirit of the 1960s space race. They found themselves involved in intense rivalry for pioneer status in studying the surface of the Moon, Mars, Venus, Mercury, the gas planets and even the Sun. As for near space, Russia resumed the deployment of its GLONASS system, which had been interrupted in 1995. Other countries pursued similar projects: Galileo (ESA), BeiDou (China), Quasi-Zenith (Japan) and IRNSS (India). The competition among these peaceful space programs was complemented by military efforts – from the George W. Bush administration's project to create a space-based anti-missile echelon to U.S. and Chinese anti-satellite weapon tests.

In the early 2010s, however, all the spacefaring nations dramatically rolled back the scale of their research programs. Experts perceived this as the end of the new space race. In fact, that race had not changed the balance of forces in space exploration in any significant way. Just like before, the spacefaring nations today fall into the following five groups:

- those that command the full range of space technologies (the United States, Russia and China);

- those that are capable of implementing individual space projects (the ESA, France, Japan and India);

- those that are building potential for future space programs (Brazil, Iran, South Korea and New Zealand);

- those that cooperate with other countries on space projects (Argentina, Venezuela, Israel, Indonesia, Vietnam, Thailand, etc.);

- those that supply individual rocket technologies to the global market (Ukraine, Belarus and Pakistan).

The six-year stagnation that followed the end of the new space race was only logical, as the spacefaring nations had realized the limits to their scientific and technical capabilities.

The United States focused on unmanned deep space studies (primarily in the form of mapping the planets of the Solar System). NASA still leads this segment of exploration, but the agency did cut back on the number of such projects in the second half of the 2010s due to their high cost and the absence of immediate return-on-investment results. The flipside of these achievements came in the form of the country falling behind in orbital missions, where its space activity became largely dependent on Russia. Despite all the declarations, the United States has so far failed to create a new generation of manned spacecraft for near-space missions.

The 2012–2018 stagnation highlighted one strategic difficulty for the United States, namely that NASA had started to gradually lose its global leadership in the field of astronautics. Just a decade ago, all the other spacefaring nations, Russia included, would set their space exploration priorities based on the NASA model (as far as their financial and scientific resources would permit). Now there is this feeling that the United States has failed to meet the goals it set itself back in 2004. President Donald Trump's December 2017 call to resume the space race has not yet produced any tangible results. [2] The United States still preserves its leadership in unmanned space flight and deep-space exploration, but other countries are increasingly beginning to identify their own space priorities irrespective of what NASA is doing. In theory, these countries may be capable of implementing breakthrough projects without any reference to the actions of the United States in future.

Russia retains its leadership in manned space flight, and in the foreseeable future Russian Soyuz spacecraft will remain the only means of delivering crew to the International Space Station. Russia commands the world's second largest satellite constellation after the United States [3]. In the spring of 2012, the country brought its GLONASS constellation up to the 1995 level. Russia effectively holds a monopoly to manned launches and leads the market for commercial launches. However, the country's current space achievements are largely a fallout of the Soviet legacy, primarily Soviet rocket technology, and the achievements of the national space agency (Roscosmos) chiefly boil down to the modernization of Soyuz manned spacecraft, upgrades to communications satellites, and to converting Strela and Rockot light launch vehicles from the UR-100N heavy ICBMs that are due to be eliminated under the START 1 Treaty of 1991. In its excessive fixation on near space, Russia found itself cut off from the study of deep space. The failures of Roscosmos' Lunar, Martian and Venerian projects have come in particular dissonance with the successes (albeit limited) of not just the United States, but even China and Japan too.

There is another problem. In the 1960s (in fact, even way back in the late 1940s), the Soviet leadership committed a number of strategic mistakes in the development of electronic computing technology. These mistakes resulted in the USSR losing to the United States, Western Europe and Japan in terms of information and communications technology. The Soviet defence industry had great potential for improving the existing weapon systems. Development programs for specialized and dual-use electronic systems were launched in the 1990s. In 2005, Russia resumed efforts to modernize these industries. However, the fundamental reasons for Russia's inferior electronic and computing technology have not yet been eliminated. Hence the current gap between Russia and the United States – and even between Russia and some EU countries – when it comes to building satellites and creating deep-space exploration systems.

China has also exhausted the space exploration potential it had back in the early 2000s. The country managed to become the third leading space power by carrying out manned space flights, testing anti-satellite weapons, creating its own orbiting station and implementing projects to launch unmanned probes to the Moon. China is working to develop the next-generation Shenlong reusable space transportation system, but it still has at least two problems. First, the country is heavily dependent on Russian space rocket technology (in fact, it was access to that technology which enabled China to make its breakthrough in space in the first place). Second, China does not have its own national school of fundamental sciences. So far, the best the country can do is match the level of achievements that the USSR and the United States reached back in the 1970s.

The second-tier spacefaring nations remain stagnant. The ESA has put its projects to develop deep space on hold and has not even completed the creation of the Galileo navigation and communications system, which was to become an equivalent of the U.S. NAVSTAR and the Russian GLONASS. [4] Although Galileo was put into operation in December 2016, it is still functioning with limitations, meaning that it cannot independently provide round-the-clock global coverage so is compensated by NAVSTAR satellite data. In 2010, Japan effectively froze its major deep-space projects and shifted the focus to launching weather satellites. India's most recent achievement was the successful launch of the Mangalyaan spacecraft into the Martian orbit in 2014, but this and other successes resulted from projects launched back in the first half of the 2000s.

In light of the above, we should not expect any serious breakthroughs in the development of space programs between now and 2024. None of the spacefaring nations (the United States included) is currently implementing any research projects comparable to the new space race of 2004–2011. Such projects may theoretically be launched in 2019–2022, but their practical implementation would not bring any results before the late 2020s. The second-tier spacefaring nations do not have the required research and technological potential to implement major space projects by themselves. In other words, all we can expect between now and 2024, for example, is for the ESA to complete the deployment of its Galileo system, China to finish work on its BeiDou system and increase the size of its satellite constellation, NASA to improve its Mars Rover technology and launch unmanned probes to the Moon, and Russia to implement its Luna 25 project to develop unmanned spacecraft capable of studying the surface of the Moon (in May 2018, the deadline for the project was once again postponed, this time until 2021).

Similar processes can be observed in military space programs. The efforts of the United States to deploy a missile defence system has thus far been limited to the creation of ground-based interceptor missiles – "extremely high anti-aircraft guns." The new projects to develop a space-based anti-missile component, which were first floated in the mid-2000s, have not yet been implemented in practice. The situation with anti-satellite weapons is no better: both Russia and the United States are still poring over research data from the 1970s, and China tested its first-generation system in 2007. The great powers putting their space rocket programs on ice suggests that they may not possess the technology for a new breakthrough in military space efforts, even though experts were convinced back in the mid-2000s that such a breakthrough was all but inevitable. Most military space projects implemented in 2009–2015 as part of the U.S. Prompt Global Strike have proved ineffective.

A possible breakthrough can only occur in the creation of microsatellites weighing in at under 100 kg. According to the United States, the official tasks of these satellites include inspection, rendezvous and docking with other spacecraft. The small-sized inspection vehicles fit into the new "swarm" military concept, which is also being implemented in other military areas in the United States. In theory, some of these new microsatellites (including the XSS 11 experimental model) could be used for intercepting the satellites of other nations under the pretext of combating space debris. To counter this threat, other spacefaring nations may actually need to develop "active systems" capable of destroying small satellites. Ground- and space-based lasers (so-called directed-energy weapons) could be used for this purpose.

However, whether or not these projects will be implemented is a moot question. Carrying them out would require certain technological breakthroughs that may not be achieved in the next six years. This is due to fundamental problems. The space breakthrough of the 1960s was made possible thanks to two factors. First, there was the widespread introduction of natural and exact science disciplines in the U.S. and Soviet primary school systems. Second, the governments of the two countries were funding major projects that did not provide immediate results. For all their outward differences, the USSR and the United States followed a fairly similar model right until the mid-1970s.

Neither of these two factors is relevant today. National governments are increasingly less capable of mobilizing resources, and the world is dominated by giant bureaucratic systems that inevitably operate to the benefit of different influential lobbies whose main interest lies in securing immediate profit. This restricts the development of long-term, expensive space exploration projects.

Another problem has to do with the changing quality of education. Experts say that over the past 30 years, the quality of natural-science teaching has dropped in the United States, Russia, and especially in the European Union. Traditional lectures and seminars are being replaced by various games and skill learning methods. The results are obvious: physicists are concerned that over the past 50 years, there have been almost no major discoveries in the domain of natural sciences anywhere in the world that would be comparable to the achievements of the first half of the 20th century.

In theory, the problem could be solved by cooperating on space exploration projects. The achievements of the United States in deep-space exploration could be combined with Russia's orbital experience. However, given the growing hostility between the two countries, this option appears to be a fantasy. A growing military rivalry in near space between the United States and Russia is more likely. Coupled with the sanctions war between the two countries, this could even affect the fate of the International Space Station, the only successful cooperation project in space that remains.
RIAC FORECAST 2019–2024
Close Your Eyes and Do It: Cybersecurity in 2018
Maria Smekalova
RIAC Website Editor, Coordinator of the Russia–U.S. Dialogue on Cybersecurity Project
Key Cybersecurity Events in 2018

2018 was rich in cyber events, offering a host of reasons for analysis and reflection before the year was out.

Cybercrime. According to Group-IB reports, the majority of cybercrimes in 2018 were committed by "pro-government" hackers, and were not financially motivated, as was the case in 2017. As a rule, pro-government hackers aim to break into critical infrastructure networks (primarily in the energy sector). One of the most serious cyberattacks last year was the Olympic Destroyer malware that shut down the official website for the Winter Olympic Games in Pyeongchang. As for attacks on the banking sector, the infamous North Korean Lazarus Group and BlackEnergy continue to operate, with the Silence hacking group is gaining influence.

In addition, Group-IB points out that Southeast Asia is the most active region in terms of cyberattacks. Despite the fact that the countries in the region have expressed serious concerns about the current situation and are taking steps to increase the level of security, the quality of dialogue in the region and the issues that need resolving are similar to the situation experienced in Western countries in 2010–2012. Given the fact that these countries have already developed mechanisms for minimizing certain vulnerabilities and are able to prevent a number of attacks, the hackers are going for easy targets in countries with less rigorous security infrastructures in place.

The Russian factor. 2018 marked the third year in a row that Russia was in the cyber news headlines. Things kicked off in 2016, with accusations of hacker attacks and interference in the U.S. presidential elections. Investigations continued in 2017, and in 2018 fuel was added to the international fire of mutual mistrust and suspicions. Continuing the topic of cybercrime, losses from hi-tech theft during the second half of 2017 and the first half of 2018 exceeded 3 billion roubles (or $53 million). The Law on the Security of Critical Information Infrastructure entered into force on January 1. In addition, a cybersecurity competence centre was set up at Rostelecom State Corporation. [1] The national operator now controls 70 per cent of the Russian cybersecurity market. The Centre's functions will include a range of areas: from monitoring and responding to cyber incidents to developing new technologies and setting specialized centres in the commercial sector and government bodies. The Bank of Russia is now responsible for cybersecurity issues in the financial sector.

Federal Law No. 374 "On the Introduction of Amendments to the Federal Law 'On Countering Terrorism' and Other Legislative Acts of the Russian Federation Related to the Fight Against Terrorism" dated July 6, 2016 (better known as the Yarovaya Law) entered into full force in the summer of 2018. According to Article 15 of the Law, internet operators are required to store data on the transmission of text, audio and video messages sent on their networks, as well as data on who sent and received these messages, during one year after such information is sent. The contents of these communications are to be stored for six months after they have been sent. Operators are further required to provide this information to the authorities upon request.

One event that caused a stir was the blocking of the Telegram messenger service, which for its part demonstrated an unprecedented ability to circumvent blocked IP addresses and continue its work. Both the authorities and the law enforcement agencies are concerned about the situation surrounding access to instant messenger services. In October, the government approved the rules for verifying users: instant messenger services are now required to check the registered phone numbers of users with mobile operators in order to confirm and verify their identities. [2] Representatives of the Federal Security Service (FSB), for their part, talk about the need for the intelligence services to monitor and control cyberspace, which will allow the country to conduct an active fight against cyber threats and terrorism. [3]

Serious progress was made last year in terms of international initiatives. Both information security documents developed and submitted by Russia went through the approval process at the United Nations General Assembly First Committee and received the support of a large number of states. [4] In December, the UN General Assembly overwhelmingly endorsed the 13 principles of peaceful behaviour of states in cyberspace proposed by Russia, marking the culmination of two decades' worth of work in this area on the part of Russia. Unsurprisingly, western states were against the initiative.

American Rides. Cybersecurity was also hot on the agenda in the United States last year, with a number of doctrines directly relating to the issue being introduced. September saw the publication of the National Cyber Strategy of the United States, one of the year's most anticipated documents (two separate documents were also released – one by the Department of Homeland Security and another by the Department of Defense). The document proposes taking a more proactive and even offensive stance on the issue, much to the satisfaction of the military. According to the document, law enforcement agencies associated with the Department of Defense will be allowed to commit offensive actions in cyberspace (for example, hacker attacks in response to the actions of other states). The presidential decree has opened up unprecedented opportunities for using cyber weapons. [5] The rationality of this step caused more disagreements among experts and the general public alike. In this context, the number of emerging threats may exceed the number of problems that need to be solved: false-flag operations will create a new round of tensions. The document also implies that the authorities will work closely with the operators of the main critical infrastructure in order to ensure the highest level of cybersecurity.

The United States has identified who has been causing the problems in cyberspace – namely North Korea, Russia, China, Iran and international terrorist groups. The highly anticipated results of the Special Counsel investigation led by Robert Mueller should provide some closure with regard to the Russia's supposed interference in U.S. politics.

The United States is ready to share its growing cyber potential (both offensive and defensive) with its NATO allies should the need arise and if it receives the relevant request from member countries. [6] NATO also plans to set up a separate centre for coordinating the organization's actions in cyberspace. [7]

Not to be outdone in terms of proposing international initiatives, the United States put its own draft UN resolution forward as an alternative to the document presented by Russia. The draft resolution was approved by the United Nations General Assembly First Committee (mostly by the United States' traditional allies). New life was thus breathed into the dialogue on how states should conduct themselves in cyberspace, with the active involvement of a number of players (even France, which had prepared its own alternative cyber code of conduct, managed to jump on the wagon). [8]

Cybersecurity 2019 and a Look into the Future

According to Group-IB forecasts, hackers are showing no signs of slowing down in 2019 and will step up their work if they are to carry out operations and cyberattacks on an even larger scale. [9] The company believes that, as a result of the active work to find and detain members of large criminal groups (primarily from North Korea: Cobalt and Anunak, among others), people affiliated with those groups will set up new hacker crews, changing countries and regions and copying each other's tactics. As of right now, specialized organizations and companies have discovered the modus operandi of key figures, meaning that this breakdown will lead to an even more difficult process of attributing attacks and activity to a given group. We can already see the beginnings of such activity – the U.S. company FireEye discovered that North Korean hackers used Russian in their code in order to confuse investigators. [10]

The cryptocurrency market is expected to be hit hard, and private houses, home routers and data storage systems are at greater risk. Group-IB identifies motherboard manufacturers and companies that provide software for the state sector (which usually have poor security systems) as potential targets for cyberattacks. CISCO sees distributed denial-of-service attacks (DDoS attacks) as being particularly dangerous, as such attacks are carried out via home devices that are connected to the internet (the Internet of Things, or IoT).

The critical infrastructure sector will increasingly be at risk in the future (seeing as attacks on such facilities require significantly more resources), as will areas that involve the use of artificial intelligence technologies. Sberbank CEO Herman Gref has warned that losses from cyberattacks could reach $8 trillion by 2022. [11]

The cyber futures of Russia and the United States. Despite the statements made by Vladimir Putin and Donald Trump on the importance of the cyber agenda in bilateral relations, we should not expect a breakthrough any time soon. The Russian side has repeatedly noted that it expects the United States to respond to the initiatives it has put forward (including within the framework of the United Nations). As far as the Ministry of Foreign Affairs of the Russian Federation is concerned, the "cyber ball" is in the United States' court. [12]

The Americans have taken a "wait-and-see" approach, as the results of the Special Counsel investigation led by Robert Mueller are due to be announced and any evidence of Russia's interference in the U.S. elections presented. Since interference of this kind is directly related to issues of cybersecurity, further steps in this area will be based on the investigation's findings. U.S. political leaders expressly avoided conversations on this topic so as not to affect the outcome of the midterm elections. Interestingly, this "wait-and-see" approach did not get in the way of a number of position papers being published.

In spring 2018, Russia announced that it would not provide Washington with any unilateral guarantees with regard to cybersecurity and non-interference in internal processes, and in elections in particular. [13] Given the current foreign political situation, it is naïve to expect the United States to put forward any initiatives that would go some way to achieving an agreement of cybersecurity. Most likely, the situation will continue to develop in its current format, or will lead to a significant level of silent distrust between the two nations.

The issue of cybersecurity has also affected UK–Russia relations. Prime Minister Theresa May threatened to use a cyber weapon as part of the United Kingdom's asymmetrical response to the actions of Russia. [14] However, the likelihood that the United Kingdom would actually consciously increase the level of tension in bilateral relations through the use of cyber tools is extremely small. The main reason why this situation came about in the first place is that it served as a convenient distraction from the fast-approaching Brexit. The United Kingdom is left with less and less room for manoeuvre in terms of adapting the economy and legal framework so that the state can continue to function and cooperate with the European Union.

Global cooperation. In 2018, Russia continued its active attempts to "hold" a global convention on international information security. It would be reasonable to expect the discussion to continue at the level of the United Nations in 2019, although we should not expect negotiations to be easy here. Firstly, because other countries have come up with alternative proposals of their own. Secondly, because Russia is still hostage to an unfavourable foreign political situation. Right now, western countries see Russia primarily as a source of threats rather than as a potential partner (things are, of course, different with its traditional allies). What is more, cyberpolicy has become a very delicate issue for each individual state, which can even be seen in the disagreements over specific details among the countries in collective West. It is only logical that the topic be promoted to the level of a global convention or agreement within the next five years. Indeed, we should expect this – it is possible that the countries will come to an agreement after further rounds of work of the United Nations Group of Governmental Experts and an even greater number of common threats are identified. It is difficult to say whether or not such an agreement will be a modified form of the proposals put forward by Russia, a variant proposed by the United States or an entirely new document. The fact that a new round of negotiations on a subject that has over the past three years been recognized at the official level as toxic will take place is in itself commendable.

We can also expect a number of states to update their existing cyber strategies, many of which were developed at a time when the internet and the associated cyberthreats were something very different to what they are today. New cyber strategies are needed in India, Japan and a whole host of countries in Southeast Asia, where work to combat cyber issues is seriously lagging behind that of the West. We cannot underestimate the influence of China on the overall agenda: western countries are already suspicious of software produced in China and its pro-government hackers. And this trend is likely to continue throughout this year.
RIAC FORECAST 2019–2024
Development Prospects for AI Technologies
Dmitry Scheftelowitsch
Research Fellow at TU Dortmund University, Specialist in Autonomous Decision-Making Systems
Developments in Recent Months

The explosive growth in the popularity of "deep learning" algorithms, made possible thanks to the equally rapid growth of available parallel computing capacity, has enabled researchers to expand the scope of pattern-recognition and inference problems that can be solved by computer. The new features make it possible to not only solve riddles such as the well-known river-crossing puzzle involving a fox, a goose and a bag of beans, but also to automatically check the functional correctness of software – both for handwriting recognition and to single out people in a crowd with a high degree of confidence.

The most prominent deployment of AI technologies in the past several months has been the social credit system in China, where all the information collected by the government's intelligence services about every single citizen is integrated within a unified database. [1] Decisions are then made on the basis of this information with regard to providing services to a given citizen or on putting them under increased security surveillance. In fact, similar scoring systems existed before, but their scope of application was normally limited to the banking sector and decisions on granting loans or leasing property. [2] Besides, they operated on a very limited amount of information about the person being vetted.

Little wonder, then, that society has now become aware of the possible socio-political implications of introducing machine-learning technologies and algorithm-based decision-making processes. The world's leading research institutes for international relations, such as Stockholm International Peace Research Institute (SIPRI), organize international conferences on the possible impact of artificial intelligence on strategic stability. [3] And activists around the globe, including prominent scholars and IT experts, campaign against the introduction of AI into military systems. [4] The problem of creating robots that would be smart but at the same time friendly has turned from an outlandish topic into a subject of mainstream research on artificial intelligence.

From the technology perspective, the issue of whether computers can be trusted has transformed into such questions such as "How can you cheat an algorithm?", "How can you prove that a specific algorithm cannot be cheated?" and "How can you prevent an algorithm from reproducing stereotypes put into it by the data?" It turns out that, at this point in time, it is easier to cheat classifiers than it is to create programs that are resistant to interference, which is something that several major research groups have been working on.

Five-Year Forecast

The development of AI technologies and their gradual introduction into all aspects of our life have generated public demand for two things: first, an explanation of the criteria that the machine uses to make decisions and classify objects on the screen in a certain way; and second, that AI algorithms operate in a stable manner despite minor changes in the input data. These are essentially two aspects of the same problem – people are not prepared to trust machines whose operating principles are often difficult to explain and whose advertised performance is even more difficult to guarantee to resolve global issues. The issue became particularly relevant following the introduction of the Chinese social credit system. To address these problems, which are largely capable of determining the future role of AI, researchers are concentrating on two related areas: eXplainable AI (XAI) and Model Checking AI. Related fundamental research has also been conducted in the domain of statistics, where one of the most popular topics at the moment has to do with searching for patterns and causal relations in observations.

XAI This concept covers research aimed at adding symbolic reasoning capability to sub-symbolic algorithms (i.e. those operating at the level of numerical manipulations and not at the level of inference). [5] Such algorithms are represented by popular classifiers like neural networks and support vector machines (SVM). Ideally, the system should work like this: the visual classifier identifies an object in a photograph as a cat not merely based on the operation of the neural network, which outputs the different values like "cat," "dog" and "hedgehog," but also on the basis of logical conclusions, such as "a fluffy, three-coloured animal with triangular ears and a long tail indicates a cat." In theory, data for such conclusions should be provided by some initial classification based on a non-symbolic analysis of the photograph. The ultimate goal is to create classifiers that could logically prove their conclusions to the user. This should make the algorithms more resistant to attempted manipulation, and also enable formal, mathematically demonstrative arguments proving the accuracy of the classification. In addition, the end user will be able to better assess the algorithm's performance if the latter provides a semantic substantiation of its decision. [6]

If this problem is solved, it will represent a step towards the creation of artificial general intelligence, at the very least in terms of the ability of AI to demonstrate human-like reasoning. At present, it is difficult to estimate the required hardware capacity for such capability to materialize.

This area of scientific research is still in its nascence, and it is as yet unclear which of the existing or incipient approaches will prove more successful and practically feasible given the current hardware capacities. It is safe to assume that one of the research groups will manage to create a prototype in the coming years. However, it remains to be seen whether that prototype will become a major breakthrough in the efforts to create a strong AI, and what difficulties it will face. It is possible that the development of operational high-performance algorithms will prove impossible.

Model Checking AI The same goal (albeit with the use of slightly different methods) is being pursued by research teams studying existing AI algorithms and looking to formally prove their properties or modify the algorithms so that these properties could be proved. The latter category includes efforts to create neural networks that would be resistant to changes in input data (classifiers for which one could prove the property "if not more than 10 per cent of pixels change in the input data, then the classification result will remain unchanged"). [7] This is an important problem, as current neural networks can be made to change an object's classification by feeding it a doctored image in which the changes are imperceptible to the human eye. The former category of research is aimed at expanding the scope of application for automated deduction methods so that they can be used in a new context. For example, for simple problems like controlling a cart with a stick placed vertically on it, it can be formally proven that the control algorithm will prevent the stick from falling. [8] However, being able to prove the properties of more complex neural networks with numerous parameters is still far off.

The notion of algorithmic fairness merits special mention. The recent case involving Amazon highlighted two facts: [9] first, that AI technologies are being actively used in business processes; and second, that their quality often directly depends on the quality of the input data. For example, if the training data shows dominance of candidates of one sex over other, then it is easy to predict what conclusion the classifier will draw from this. In the future, it is likely that classification algorithms will be required to meet certain standards of "fairness," possibly at the legislative level. The situation will depend on the public demand for such fairness, and it may differ from country to country: for example, it is likely that legislative requirements for AI used in hiring stiff and for similar purposes will be introduced in Europe, but not in developing countries.

The growth in computing power will likely help researchers prove increasingly more complex properties of increasingly more complex algorithms. It is also likely that an algorithmically effective resistant deep learning model will be created in the coming years: up to now, all attempts to develop such a resistant algorithm have run into the high algorithmic complexity of the software, which is much more time-consuming than its "non-resistant" counterparts. [10] Under the optimistic scenario, resistant algorithms will not only be created in university laboratories, but will also be promptly commercialized; this depends, to an extent, on public and legislative pressure for such a development.

New Possibilities

The most obvious development in AI will come in the form of the new possible social and economic applications of this technology. Automation has long been used successfully in searching for anomalies in audit data, and it would be logical to see these same methods embraced by law-enforcement agencies, which have been actively introducing machine learning in crime prevention. [11] Previous attempts to introduce such methods did not bring the expected results, [12] but the next generation of models may perform better, given that the militaries of different countries are attempting to use AI to predicting the intentions of their adversaries.

As AI gets better at solving increasingly more general problems, we may start to see achievements similar to those associated with algorithms for solving logical formulas and integer linear optimization with regard to popular AI concepts. It is thus quite possible that neural networks will be used where exact computation is too expensive or where it is difficult to model the problem. For example, the problem of recognizing obstacles in real-life situations: pollution and natural and artificial interference make it difficult to model the problem (not to mention solving it), which means this could be just the task for a neural network.

The aforementioned idea is not merely of theoretical interest. It is also central to the depth reinforcement learning algorithm, which, instead of using analysis in modelling the behaviour of a complex system, employs a neural network. [13] That network is presented with a sufficient number of examples and is tasked with understanding the structural properties of the particular system (say, the strategy board game Go) and extrapolate this knowledge to unfamiliar situations. There is every reason to believe that deep training can be used not only for beating a human opponent at Go, but also for calculating and optimizing the control parameters of complex technical systems. However, some systems may prove too complex for neural networks, in which case much more time and energy will need to be spent on solving their control problems.

Of special interest is the possibility of using AI technologies to better understand the human mechanisms of perception and behaviour. Conceptually speaking, even a failure to create a machine that would think like humans helps us to understand the way we think by answering the question of what is lacking. In terms of more practical applications, studying the characteristics of human perception facilitates the creation of audio-visual effects that humans cannot distinguish from reality. The entertainment industry is very much interested in new technologies that already make it possible to create lifelike videos where the facial expressions of one person can transferred to the image of any other person. [14]

The Prospects of Strong AI

In conclusion, let us consider the prospects of the most interesting and controversial topic: the creation of a strong AI, or "superintelligence." In many respects, the possibility of a strong artificial intelligence is closely linked to the question of what intelligence actually is. At present, many components of what can be described as intelligence are already available: there are algorithms for spatial orientation, [15] partially observable decision processes, [16] and for pattern recognition in images and data. So what is missing? We believe five important components are still not there: autonomous follow-up training in real-life situations; the integration of the known methods into a single whole (even though Boston Dynamics [17] is working to bridge this gap); computing power or high-performance algorithms for shorter cycles of observation and action; a formulated target for an artificial general intelligence; and an understanding of why AGI is needed. In order to solve the last two problems, researchers need to answer all the questions that philosophers have been asking throughout the history of humanity. In the worst-case scenario, however, it may not be possible to solve the philosophical problems; this will happen if computational capacities hit a wall and a strong AI turns out to be too slow.
RIAC FORECAST 2019–2024
Short- and Medium-Term Global Migration Forecast
Dmitry Poletaev
Ph.D. (Economics), Director of the Centre for Migration Studies, Senior Research Fellow at the Institute of Economic Forecasting of the Russian Academy of Sciences, RIAC Expert
The World in 2019

Right wing parties around the world, and all conservative parties that call for migration restrictions to be adopted in developed countries, will continue to bolster their positions in 2019. Regardless of the success of these parties and the role they play in national parliaments and governments, the overall degree of apprehension towards refugees and undocumented labour migrants will continue to grow. In this regard, new initiatives to protect borders may very well be proposed by both the national governments of developed countries and inter-state associations (the European Union, for example). Despite the optimism surrounding the signing of the Global Compact for Safe, Orderly and Regular Migration (GCM) in Marrakesh in December, 2018 this will not stop countries that are attractive to migrants from forming new security systems on their borders and transitioning to more selective forms of accepting and selecting both migrant workers and those seeking permanent residence. As a result, there is a risk that networks of shadow intermediaries that facilitate the illegal transportation of migrants to Western Europe and the United States could start to gain power and expand their operations, and that the number of undocumented migrants in these countries will grow as a consequence. The shadow economy may also grow due to the unofficial employment of migrants who have crossed the border illegally.

The Russian Federation in 2019

Russia is currently witnessing a sharp reduction in the working-age population. It is thus extremely likely that fresh measures will be taken to remove barriers to labour. For example, cooperation with Uzbekistan (Russia's largest donor of labour power) will increase. The two countries are expected to step up their cooperation on migration, and barriers for labour migrants moving from Uzbekistan to Russia may be removed.

Student migration to Russia is likely to expand due to the influx of foreign students to technical and vocational education and training (TVET) institutions. This focus has already been included in the latest concept of migration policy. That being said, the scale of this expansion will not be great, as Russian TVET institutions are ill prepared to take on high numbers of foreign students.

The introduction of new measures aimed at tightening control over the process of registering foreign nationals at their place of residence in the summer of 2018 will have a number of negative consequences in 2019:

- corruption risks associated with the sale and purchase of falsified registration stamps may increase (landlords often shy away from registering labour migrants when renting housing to them);

- it will be harder for migrants to secure places for their children in kindergartens and schools, since electronic registration of pupils cannot be completed without a valid registration stamp, and if parents attempt to use a fake registration stamp, this will be detected immediately upon cross-checking with the relevant database;

- the chances to obtain Russian citizenship will decrease for migrants who, for whatever reason, have used the services of shadow intermediaries to get fake registration stamps.

Despite all the negative effects of migration, which have not been offset by the various adaptation and integration programs set up in Russia (and are not due to receive any funds from the budget in 2018), the risks associated with migration (terrorism, the "brain drain," growing xenophobia, and fear of migrants) are not particularly relevant in 2019.

The World up Until 2024

Due to the increasing digitalization of the global economy, the most intense competition in both the United States and the European Union will be for IT and computer technology specialists.

Environmental migration will become more widespread, and there will be increasing numbers of refugees flowing in from countries with ecological (and thus economic) problems – people trying to make their way to developed and/or neighbouring countries. We are talking here about countries that will be affected by rising sea levels (for example, Bangladesh), as well as countries in Central Asia that have an arid climate.

If the policies of international organizations that deal with various aspects of migration processes (the International Organization for Migration, the International Labour Organization and the Office of the United Nations High Commissioner for Refugees) develop in the same vein that they are today, then their role and influence will continue to fall.

On the other hand, NGOs, which have increasingly close ties with national governments, are globalizing and creating their own networks for carrying out their activities at home and on the international stage. They will become serious actors both in terms of providing direct assistance to different categories of migrants and in terms of advising local and regional authorities, national governments, and intergovernmental associations.

There will be a slowdown in the number of international students coming to Russia compared with previous decades. On the one hand, India and China — the largest suppliers of international students in the world — have attracted huge investments into their universities and dramatically improved the quality of education they provide. This is why the middle class in these countries prefer to have their children educated at home. On the other hand, the United States and Australia, and more recently the Netherlands, Sweden and Switzerland, have cut back on their efforts to attract foreign students due to the more conservative views on migration that have appeared in these countries in recent years. In the United Kingdom, the flow of foreign students into the country (primarily from the EU) has been interrupted by Brexit. In short, the traditional suppliers of foreign students have improved their own higher education institutions, while the traditional education exporters have shifted their priorities away from attracting students from overseas.

European countries face an imminent labour shortage as a result of the aging population, so competition for jobs, particularly those that require specialized qualifications, will gradually increase. On the other hand, the fact that the population is getting older will affect the care services, and we will see a growth in the number of domestic workers. In the event that there is competition for skilled workers between migrants from countries outside Europe and local specialists from Eastern and Western Europe, then, in the case of domestic workers, those from outside Europe will almost certainly keep their jobs. This is confirmed by the experience of immigrants from Moldova and Ukraine, who have successfully been relocating to Western Europe for decades now.

In the medium to long term, migrants from Africa and the Greater Middle East will continue to try and resettle in Europe. The active hostilities that have plagued the Greater Middle East in recent decades are largely in the past now, but the economies of the region's countries have suffered terribly, and displaced persons have little interest in returning to their home countries in their current state. Family reunification is also a reason why people are abandoning the Greater Middle East for Europe.

The flow of migrants from Africa carries a number of special risks. The high birth rate in these countries, coupled with the limited employment opportunities at home, the dim outlook for economic growth and the constantly improving transport infrastructure are all factors that contribute to the outflow of migrants from Africa and into Europe. As of 2009, the continent of Africa had a total population of 1 billion people. The number is expected to hit 2 billion by 2040, and it will continue to grow. African migrants, just like those from the Greater Middle East, typically head for Europe, and the restrictive measures being implemented by European countries will eventually lead to the increased terrorist threat, as not all migrants are able to legally resettle in the new country, find long-term official employment and live at a decent level compared to the native population. Some of them will be at risk and vulnerable to recruitment by radical groups and terrorist organizations.

European countries will not be able to effectively counteract this uncontrolled movement, but they can prepare for it by developing and improving migration management plans in order to mitigate the negative effects as much as possible. We are already seeing a sharp increase in funding to protect EU borders, as well as a heightened interest in the development of a readmission capacity building facility that would serve to return migrants to their homes. These practices will be further developed and improved.

At the same time, the development assistance programs created by first world countries to assist vulnerable states will not be effective enough to resolve the issue of permanently replenishing the latter with migrants.

The growing presence of right-wing parties in the power structures of Western European countries in recent years allows us to predict with a high degree of certainty that EU governments will gradually expand their powers in regulating migration flows and reducing the legal possibilities for migrant workers from non-EU countries, and possibly from Eastern European countries that have already joined the EU, to cross their borders. This state of affairs will lead to a gradual increase in the number of undocumented migrants and the disappearance of the niches of the economy that currently attract migrant workers into the "shadows": construction, domestic jobs, agriculture, the hospitality business, catering, street cleaning, office cleaning, etc. [1] The "right movement" can also be observed in the United States, and will have to be dealt with in the medium term.

The Russian Federation up until 2024

The demand for foreign labour in Russia will lead to the gradual loosening of the requirements for foreign labour migrants. The 2018 concept of the migration policy of the Russian Federation until 2025 provides convincing proof of this. While introduction of the licensing system as a tool for legalizing the employment of workers from CIS countries may be considered a form of weakening control over labour migration, the rules with regard to foreign workers from EAEU countries are even looser, as they do not even need to obtain such a document to legally work in the Russian Federation. The conditions that make it difficult for foreign workers to enter the Russian labour market will be further simplified in the future.

Nevertheless, Russia has adopted a controlling rather than an economic approach to managing migration processes in recent years, although it has weakened over time. At the same time, the current situation is largely being preserved as a result of the significant shadow economy and corruption in the field of migration. The tightening of the registration system (which has long been ineffective in terms of managing migration in Russia) means that corruption in this sphere will continue to grow and legalizing undocumented migrant workers in Russia will become even more difficult.

Thus, even in the framework of the Eurasian Economic Union (EAEU), we are seeing an increase in shadow relations and the expansion of the grey area of the Russian economy. While immigrants from Kyrgyzstan are no longer required to obtain a license to work in Russia, they still use the services of intermediaries to purchase fake registration certificates at their place of residence, as landlords refuse to register them.

The risk of tuberculosis and HIV spreading among EAEU migrant workers in Russia will increase in the medium term, as the removal of the licensing system means looser control over these dangerous diseases among this section of the population.

The current level of corruption in the migration sphere will, in the medium term, further complicate the manageability of migration processes and preserve the shadow system of employment of foreign workers. This, in turn, will bring about a transition on the part of migrant workers from adapting to the conditions of informal work to the consolidation and institutionalization of already shady migration practices.

The lack of integration programs, an official specialized infrastructure and services for different categories of migrants (primarily for migrant workers) increases the risk of social discontent in Russia and could lead to migrant communities banding together, none of which is conducive to the unity of Russian society.

Just like Europe, Russia should expect an increase in the threat of terrorism in the medium term, as a significant number of migrants in the country are in a difficult situation and their working conditions are extremely challenging.

Two kinds of radicalism may emerge in the long term (8–10 years): 1) the growing radicalization of Russian nationalists; and 2) the radicalization of migrant workers and Russian citizens of foreign origin, including children of migrants who have grown up in Russia and are seen as particularly vulnerable by radical (including religious) groups. And these risks are very real, a fact that many Russians have witnessed first-hand. Examples include demonstrations on Manezhnaya Square in Moscow, the "migrant factor" being used as a reason for the Biryulyovo riots, the terrorist attack on the St. Petersburg Metro, the recruitment of radicals from among the migrant community by Islamic State (a terrorist organization that is banned in Russia), the emergence of "prison jamaats" that foreign citizens serving time in Russian correctional institutions join, etc. Such cases are rare, but they risk becoming more frequent moving forward. Having said this, these risks are unlikely to materialize in the medium term. An important factor here is the fact that Russia is a country of migrants. It has a rich migrant history, and this gives it a certain margin of safety, immunity for the time being.

As is the case with other countries in Europe, the gradual aging of the population in Russia will lead to the expansion of the care services. Selective studies carried out over the past five to seven years have demonstrated this growth. Right now, migrant workers compete with locals (who are often migrants themselves, but internal migrants) for jobs in this sector. Having said this, the competition for niche positions that involve caring for the sick and elderly is not especially high among Russians, as locals a reluctant to agree to such a difficult job.

The growing need to reindustrialize Russia in the coming years to cope with the falling resources of young people will lead to an even greater shortage of qualified and unqualified workers and specialists. There are a number of projects and development areas (Akademgorodok 2.0, the modernization and development of the military-industrial complex and the further development of the agro-industrial complex) that will require the training of additional qualified workers for the Russian labour market). On the other hand, the ongoing sanctions policy and the issues being experienced by small and medium-sized enterprises will mean that qualified personnel will continue to leave the country in the medium term. Young people will look abroad for the best opportunities for growth and development, at least as a temporary measure.

The process of bringing in migrants from former Soviet countries to support Russia's demographic and economic potential, which partially compensates for the "brain drain," will continue. There has been a certain growth in the number of foreign students from EAEU countries in recent years. [2] Unfortunately, as we have already noted, we should not expect a significant influx of foreign students into technical and vocational education and training institutions, as they are not prepared to take on high numbers of students from abroad. The current limited flow of students can only be expanded by introducing a large-scale program to modernize TVETs.

Given the structure of the flow of foreign students into Russia, it is strategically important to step up cooperation with Uzbekistan, which was the first to make moves in this area. However, the importance of Uzbekistan as a resource will fall sharply after 2024. Environmental migration from Central Asia (including Uzbekistan) will gradually acquire an increasingly important role.

Research shows that more and more people are emigrating from Siberia and the Russian Far East to the European part of the country, which thus increases the regional imbalance in the distribution of labour. In the medium term in order to stop the population outflow from central Russia, it is necessary to search for resources and opportunities to both attract foreign labour to Siberia and the Russian Far East and improve the conditions for Russian people living in these regions.
1. Moltz J. The Politics of Space Security. Strategic Restraint and the Pursuit of National Interests // Stanford: Stanford University Press, 2008; Freeze J. Space as a Strategic Asset // New York: Columbia University Press, 2007.
2. Trump Signs Decree Launching New U.S. Lunar Program // RBC, 12.12.2017. URL: https://www.rbc.ru/technology_and_media/12/12/2017/5a2f0c5b9a7947eed2c43791
3. Viktor Sokirko. Reading US Officers' Badges from Space: Why the Pentagon is Panicked // Leonids-Info, 12.04.2017. URL: http://www.leonids-info.ru/analitika/orbitalnye-gruppirovki-ka-rossii-i-ssha-kto-zhe-sejchas-glavnyj-po-tarelochkam-nad-zemlej/php
4. Europe's Galileo Navigation System goes Live with Initial Operational Services // Spaceflight 101, 15.12.2016. URL: http://spaceflight101.com/galileo-navigation-system-goes-live/
1. Hunting for Hackers // Rossiyskaya Gazeta. 22.05.2018. URL: https://rg.ru/2018/05/22/v-rossii-poiavilsia-centr-kompetencij-kiberbezopasnosti.html
2. Decree No. 1279 of the Government of the Russian Federation "On Approving the Rules for Identifying Users of the Internet Information and Telecommunications Network by the Organizer of the Instant Messaging Service" // Official Legal Information Portal, 06.11.2018. URL: http://publication.pravo.gov.ru/Document/View/0001201811060001
3. FSB Says Intelligence Services Need to be Able to Control Cyberspace // RBC, 18.10.2018. URL: https://www.rbc.ru/rbcfreenews/5bc875919a79471477ba0722
4. Russia Opens Hunting Season at UN // Kommersant, 17.11.2018. URL: https://www.kommersant.ru/doc/3803933
5. Trump Looses Secretive Restraints on Ordering Cyberattacks // The New York Times, 20.09.2018. URL: https://www.nytimes.com/2018/09/20/us/politics/trump-cyberattacks-orders.html
6. United States Announces Readiness to Use Cyber Potential against Russia // RBC, 03.10.2018. URL: https://www.rbc.ru/politics/03/10/2018/5bb459ad9a7947171372a88b
7. Cyber Defence // NATO, 16.07.2018. URL: https://www.nato.int/cps/en/natohq/topics_78170.htm
8. Cybersecurity: Paris Call of 12 November 2018 for Trust and Security in Cyberspace // France Diplomatie, 2018. URL: https://www.diplomatie.gouv.fr/en/french-foreign-policy/digital-diplomacy/france-and-cyber-security/article/cybersecurity-paris-call-of-12-november-2018-for-trust-and-security-in
9. Group-IB Publishes Report on Cyber Crime and Urges the Market to Go Hunting // Group-IB, 09.10.2018. URL: https://www.group-ib.ru/media/hi-tech-crime-trends-2018/
10. North Korean Hackers Suspected of Disguising Themselves as Russians // RBC, 04.10.2018. URL: https://www.rbc.ru/politics/04/10/2018/5bb4ebe19a7947423aab5620
11. Gref Predicts Global Losses from Cyberattacks Will Reach $8 // RBC, 06.07.2018. URL: https://www.rbc.ru/technology_and_media/06/07/2018/5b3f2f2b9a79473dfc957b1e
12. Briefing by Deputy Director of the Information and Press Department of the Ministry of Foreign Affairs of the Russian Federation A. Kozhin, Moscow, July 26, 2018 // Ministry of Foreign Affairs of the Russian Federation, 26.07.2018. URL: http://www.mid.ru/ru/foreign_policy/news/-/asset_publisher/cKNonkJE02Bw/content/id/3303792
13. Ministry of Foreign Affairs of the Russian Federation: Russia Will Not Give U.S. Unilateral Guarantees on Cybersecurity // Kommersant, 22.04.2018. URL: https://www.kommersant.ru/doc/3612180
14. Times: United Kingdom Exploring the Possibility of Carrying Out Cyber Attack against Russian // TASS, 13.03.2018. URL: https://tass.ru/mezhdunarodnaya-panorama/5024495
1. Liang F. Constructing a Data-Driven Society: China's Social Credit System as a State Surveillance Infrastructure // Policy & Internet, 02.08.2018.
2. Han L., Zhao H. Credit Scoring Model Hybridizing Artificial Intelligence with Logistic Regression // JNW, 2013. № 1(8). C. 253–261.
3. Markotkin N. Close Encounters of the Third Millennium // RIAC, 28.09.2018. URL: http://russiancouncil.ru/analytics-and-comments/analytics/blizkie-kontakty-tretego-tysyacheletiya/
4. See: Campaign to Stop Killer Robots. URL: https://www.stopkillerrobots.org/about/
5. Clark K. A Framework for Integrating Symbolic and Sub-symbolic Representations IJCAI'16 . New York: AAAI Press, 2016. P. 2486–2492.
6. Holzinger A. Current Advances, Trends and Challenges of Machine Learning and Knowledge Extraction: From Machine Learning to Explainable AI Lecture Notes in Computer Science / Ed. A. Holzinger. Hamburg: Springer, 2018. P. 1–8.
7. Ruan W., Huang X., Kwiatkowska M. Reachability Analysis of Deep Neural Networks with Provable Guarantees / Ed. J. Lang. IJCAI, 2018. P. 2651–2659; Schott L. Towards the first adversarially robust neural network model on MNIST / New York: CoRR, 2018.
8. Scheibler K. Towards Verification of Artificial Neural Networks / Eds. Heinkel U, Kriesten D., Rößler M. Dresden: Sächsische Landesbibliothek, 2015. P. 30–40.
9. Reuters: Amazon Develops AI Recruiting Tool, Scraps it Over Discrimination of Women // VC, 10.10.2018. URL: https://vc.ru/hr/47756-reuters-amazon-razrabotala-ii-dlya-poiska-sotrudnikov-i-otkazalas-ot-nego-iz-za-diskriminacii-zhenshchin
10. Madry A. Towards Deep Learning Models Resistant to Adversarial Attacks / New York: CoRR, 2018.
11. Chandola V., Banerjee A., Kumar V. Anomaly Detection: A Survey // ACM Computing Surveys, 2009. № 3 (41).
12. Saunders J., Hunt P., Hollywood J.S. Predictions put into practice: a quasi-experimental evaluation of Chicago's predictive policing pilot // Journal of Experimental Criminology, 2016. № 3(12).
13. Mousavi S., Schukat M., Howley E. Deep Reinforcement Learning: An Overview / Eds. Bi Y., Kapoor S., Bhatia R., Cham: Springer International Publishing, 2018. P. 426–440.
14. Thies J. Face2Face: Real-time Face Capture and Reenactment of RGB Videos // Stanford University, 2016. URL: https://web.stanford.edu/~zollhoef/papers/CVPR2016_Face2Face/paper.pdf
15. Cadena C. Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age // IEEE Transactions on Robotics, 2016. №6 (32). P. 1309–1332.
16. Kaelbling L., Littman M., Cassandra A. Planning and acting in partially observable stochastic domains // Artificial Intelligence, 1998. №1 (101). P. 99–134.
1. Depending on the specifics of the country in question, spheres such as housing and communal services may also increase the shadow component.
2. The number of students from Kazakhstan coming to study in Russia has increased twofold since the inception of the EAEU.