On 12 November 2020, a panel discussion titled “Artificial Intelligence and International Psychological Security: Theoretical and Practical Implications” was held at St. Petersburg State University, and broadcasted on Zoom, as part of the international conference “Strategic Communications in Business and Politics” (STRATCOM-2020)[1].

Members of the Research MUAI Group from Russia, Romania and France actively participated in the panel, along with some 40 participants.[2]

The first paper “Artificial Intelligence in the Context of the Ensuring of the International Psychological Security” was presented at the panel session by Prof. Konstantin Pantserev.

The necessity of the development of advanced technologies is considered as the indispensable condition for the ensuring of the global leadership for every country in the contemporary world. Special attention is paid to technologies which are based on artificial intelligence (AI). Very often domestic developers of AI-based technological solutions are supported by governments of their countries of origin. According to official data, nowadays more than 30 countries have elaborated their national strategies and roadmaps at the field of AI. But it becomes evident that all technological novelties which should simplify our life can be used maliciously in the future. Thus the rapid growth of our dependence on hybrid computer intelligent systems makes us extremely vulnerable for plotters who use AI-based technologies both for satisfaction of their personal demands or, what is much worse, to damage the critical infrastructure of the country. This last point poses a serious challenge for international psychological security.

Konstantin Pantserev underlined following key threats of the malicious use of AI:

• The loss of control over autonomous weapon. That means once autonomous armament systems appear which will be programmed for killing. Nowadays all leading nations pay increased attention to the elaboration of different intelligent armament systems. Thus we can predict that in the nearest future the nuclear arms race will be replaced by the race of the development of military hybrid intelligent systems. But we still can’t predict what will happen if the nation loses control over such intelligent systems or if such systems would be in the hands of terrorists.

• Social manipulative practices. Nowadays social media with the aid of AI-algorithms are very effective in the field of targeted marketing. But this technology can be also used maliciously. AI-algorithms by having access to personal data of millions of people and knowing their needs, strengths and weaknesses can suggest target propaganda and manipulative practices aimed at concrete individuals.

• Invasion of privacy. This threat has already become a reality. Now a plotter can track and analyze every step of an online user, as well as the time when he or she is doing his or her daily tasks. Cameras are almost everywhere, and facial recognition algorithms easily identify us.

• Mistakes of operators. The human element will rest crucial even if smart machines emerge which can learn by themselves. AI is valuable to us primarily because of its high performance and efficiency. However, if we do not clearly define a task for an AI-system, then its optimal execution can have dangerous consequences.

• Lack of data. As is known, all AI-algorithms are based on the processing of data and information. The more data that has been uploaded into the system the more accurate will be the result. But if there will not be enough data for the achievement of the proper result or the data will be poisoned by terrorists for example – this can cause a glitch in the functioning of the whole AI-system and has unpredictable consequences for people.

Finally, Konstantin Pantserev pointed out that there is one more threat that should be considered as the most significant one. It is our increased dependence on advanced technologies that have already penetrated the everyday life of every person and become responsible for the functioning of numerous applications and even critical infrastructure. It becomes evident that such technologies will be valuable for plotters and terrorists of any type.

Prof. Evgeny Pashentsev presented his paper “Artificial General Intelligence and Superintelligence: Threats to International Psychological Security in the Process of Working on Their Creation” at the Plenary Session, which he co-chaired, then he continued this topic at the first session of the panel.

Artificial Narrow Intelligence (ANI) is the only form of artificial intelligence that humanity has achieved so far. This is AI that is good at performing a single task, such as playing chess or Go, or making purchase suggestions, sales predictions and weather forecasts. The possible creation of human-equivalent Artificial General Intelligence (AGI) in the current century, which many experts in AI pay attention to, and consequently the almost inevitable and very rapid arrival of Artificial Super Intelligence (ASI) and potential singularity, will inevitably introduce fundamentally different realities. Some aspects of the creation of AGI seem important in the context of the topic of the current panel.

The threats from AGI to the psychological security of human society appear just now when it is not created yet. Among them Evgeny Pashentsev mentioned:

• The threat of loss of meaning in life by people, the real threat of unemployment because of ANI and especially AGI, other consequences of rapid implementation of AI constitute a real threat of destabilization of public consciousness long before the real appearance of AGI.

• Explosive nature of psychological destabilization with rapid AGI progress.

• The rivalry of great powers and different other state and non-state actors with selfish interests in an increasingly socially and politically polarised world which is accompanied by sharp psychological warfare. The image of AGI will inevitably be used by the opposing sides in this confrontation for a variety of purposes. The use of AGI will be used, for example, to intimidate people: "AGI is better than us”, “we will not have a place on earth” and thus to provoke dissatisfaction with technological progress in a particular country, etc.

•  Exaggeration and understatement of the possibility and significance of creating AGI in the current century. The second is more dangerous, because it disorients and does not prepare people for possible drastic changes in their fate.

• We can consider the nature of AGI as the possibility of the emergence of an integrated mind with its will and feelings. However, its birth and initial development will take place in the human environment, based on human information and knowledge. It is important that AGI will not be a product of humanity in general, but specific people. There are also possible options, until AGI’s appearance in the laboratory, controlled by antisocial, reactionary, militaristic circles. In addition, if the environment often deforms people (of different intelligence), then why is this not applicable to General AI? Another thing is if we get an integrated powerful intellectual potential, capable of solving problems only on human instructions. Then we are dealing simply with a more powerful machine, and the pros/cons of using it will depend on the people running it. Perhaps the second will precede the first. Let’s see.

This is only part of the obvious moments that do not allow us to bow our heads under the knife of the ruthless guillotine of singularity with mystical horror. In addition, today everything still depends on people who, alas, are divided and do not think in the majority about a strategy for the development of society. The development of humanity on a progressive basis with the elimination of acute social contradictions and inequality, with the development of physical and creative abilities of a person in symbiosis with improving artificial intelligence, will allow us to move to a qualitatively new phase of human society.

Darya Bazarkina, DSc in Political Sciences, Professor at the Russian Presidential Academy of National Economy and Public Administration and Senior Researcher at the Saint Petersburg State University, presented the paper “Appeals to the Topic of Artificial Intelligence in Terrorist Propaganda: the Target Audience and Ways to Influence It (on the Example of the Magazine 'Kybernetiq').”

As she stressed, Kybernetiq magazine can be considered an example of how terrorist propaganda adapts to changing social, political, military, and technical realities. The magazine appeared four years ago (the first issue was published on December 28, 2015, the second and third ones in November 2016 and December 2017, respectively). The reactionary political actors fully understand that effective propaganda should take into account the real situation. Kybernetiq is a good example of adaptation of such propaganda to a new round of technical progress, involving new types of weapons and requiring a certain rationality of implementor’s thinking (of course, only at the tactical level).

In addition to the obvious target audience of the magazine (terrorist fighters seeking to hide their digital footprint), Darya Bazarkina distinguishes the following “risk groups”:

• Young people who are passionate about technology (as well as the cyberpunk subgenre in the sci-fi literature and art) initially. Recruiters can start a “hunt” for the part of this group, especially for the people engaged in writing their own programs, modification of computer games, etc., for example, students of faculties and departments of computer science, and separately for propaganda purposes— for young people who study design or are fond of it.

• ICT professionals. These are likely to be much rarer cases, since for established professionals the salary will be a decisive factor, while terrorists usually use propaganda to reduce the cost of recruitment. A separate category can be composed of persons who already sympathize with terrorist organizations in real life and seek to strengthen them technically.

• A wide audience of people interested in both political issues and science fiction literature, advanced technologies (a part of these people can unwittingly become new conductors of terrorist propaganda).

The use of advanced technologies’ images in terrorist propaganda creates new risks. However, by carefully studying them and having (in most cases) much greater technical capabilities, as well as cooperating with each other, state authorities, intelligence agencies, public institutions and the private sector can greatly anticipate the effect of terrorist propaganda and at least prevent recruitment. When more people are using darknet, there is a task for the technical experts to develop the tools for monitoring the hidden part of the Internet (ideally—predictive analysis tools that can work in the darknet). While the technological equipment of terrorist organizations is still overestimated, forecasting the threats associated with the increasingly widespread advanced technologies in a world where the eradication of terrorism is still far away is a matter of international security. Today, more than ever before, we have no right to prepare to fight the last war.

Darya O. Matiashova, B.Sc. (International Relations), Master student at the St. Petersburg State University, in her paper “Strategies of AI promotion in German and Indian economies: comparison and prospects for world policy system” analyzed trends and problems of strategic planning in the field of AI.

AI is a technology that can increase the productivity of all sectors of the economy, which is an incentive for its active implementation. At the same time, potential threats of both socio-economic and technological nature create a special need for principles regulating the use of AI. The successful development of such principles at the state level can bring both economic and political dividends.

In the context of India and Germany, the convergence factors are the social orientation of AI as a tool (both countries declare "using AI for the common good" as their strategic goal) taking the levels of economic and social development of these countries into account, focus on smart cities and the transport sector as on one of the main drivers of AI implementation in the economy, focus on the use of AI to promptly inform the population about emergency situations, the operation of complex monitoring systems, and the elimination of the consequences of man-made and natural disasters and, finally, perspective on the AI implementation to the economy as the result of a multi-stakeholder dialogue between research centers, the state, the financial sector, and industry. There are also such factors of the underestimation of AI threats and the general optimism regarding the AI implementation into sectors connected with high security and health risks (such as medicine, transport, rescue services) – nevertheless, they can be regarded as the convergence factors only in the long term outlook and only in case of common cooperation to prevent threats underestimated.

The divergence factors have tactical dimensions (while for India the dialogue has primarily a technocratic character, the German strategy involves a broader review of the legal framework and the introduction of public and private audit), subjective dimensions (if Germany is focused on creating common research institutes and industrial clusters with the EU, then India plans to create them together with mainly American and British TNCs and universities) and strategic ones (if India's ambitions are focused on overcoming existing economic problems and offering effective solutions to the developing world, then Germany's goals are to make German AI a global quality mark, and European AI a global industry leader). The latter can provoke intensive technological competition which is able to spill over into “standards competition” and even a trade war. The identity basis for this competition, can be rooted, in turn, in the perception of the Western technologies promotion as a “rival” and “neocolonial” one, on the one hand, and in the willingness to push the weakening economic growth in the West due to the demand in the East, on the other.

Thus, as Darya Matiashova concluded, the German and Indian principles of implementing AI in the economy bring together social orientation and focus on the transport, security and education sectors, which creates a space for cooperation and dialogue. Both states focus on the same fields where common or close industrial standards can be developed. What is more, the common values of guaranteeing “soft” security can be a factor of rapprochement. However, different views on the role of non-state institutions in shaping norms and standards, different vectors of cooperation, and Germany's global ambitions create the potential for diverging standards at the level of individual industries and, as a result, make conflicts over these standards possible.

Dmitrii Rushchin, PhD in History, Associate Professor at the St. Petersburg State University, points out in his paper “Problems and Prospects of Artificial Intelligence Development in Russia” that in recent years, much attention has been paid to AI technologies in Russia.

In October 2019 the "National Strategy for the Development of Artificial Intelligence for the period up to 2030" was approved. The National Strategy for the Development of AI defines two reference points for the development of artificial intelligence in Russia. These are 2024 and 2030. It is expected that by the first date, the country will significantly improve its position in this area, and by 2030 it will eliminate the lag behind well-developed countries, as well as achieve world leadership in certain areas related to artificial intelligence.

To understand the current level of development of artificial intelligence technologies in Russia, you should look at the selected indicators of technological development associated with it. For example, in terms of the number of supercomputers in the top 500 most powerful in the world, Russia is currently in 15th place in the global ranking. At the same time, it has only three such supercomputers at its disposal. For comparison, there are 228 of them in China, 117 in the US, and 29 in Japan.

Another important indicator is the number of scientific publications on artificial intelligence. According to the Scientific Journal Rankings, Russia is ranked 31st in this indicator.

In the field of computer science, which includes machine learning, 16 Russian universities were among the 684 best universities according to the World University Rankings 2019. However, only two Russian universities — ITMO in St. Petersburg and Lomonosov Moscow State University - were in the top 100.

One of the key indicators of the development of artificial intelligence in a particular country is the number of startups engaged in the development of these technologies. According to TRACXN company, there are currently 168 startups in the field of artificial intelligence in Russia. For comparison, TRACXN counted 6,903 artificial intelligence-related startups in the United States and 1,013 in China. TRACXN is an analytics firm co-founded by Sequoia and Accel alumni to track data from startups.

In part, the small number of startups in the field of artificial intelligence is due to the fact that in Russia, this industry is dominated by already established companies. This feature distinguishes it from most other well-developed technological powers. According to the Project named “Map of artificial intelligence of Russia”, there are a total of 400 companies engaged in development in the field of artificial intelligence in Russia.

As for the use of artificial intelligence and its level of development in the military sphere, it should be noted that due to the secrecy of data, international ratings do not take into account military developments. Meanwhile, for many decades, it was military science and the defense industry that were the locomotives of the country's technological development.

For the country's leadership, a high priority for the development of artificial intelligence is determined by significant –by Russian standards– financing of industry projects. However, Russia can hardly claim a good global position in the field of artificial intelligence with the existing level of funding for R&D (Research and Development): 2 times less percentage of GDP than in France or Singapore; 3 times less percentage of GDP than in Finland or Japan; 4 times less percentage of GDP than in South Korea or Israel.

If there is political will and sufficient funding, Russia can become a major player in the field of artificial intelligence and achieve leadership in certain areas, concluded Dmitrii Rushchin.

Konstantin Golubev, PhD in History, Associate Professor at the St. Petersburg State University, presented the paper “Oppositional News Channels on Youtube: Threat to Information Security”. He highlighted the public control over the national information infrastructure as the most important strategic objective for the establishment. A loss (partial or complete) of it would create gaps in the production of hegemonic discourse within a single national-linguistic community, hindering its ability to monopolise news agenda or to determine the framing of the most important events. Over the past few years in Russia, the development of information and communication technologies and their growing mass outreach, on the one hand, while on the other, the decline of people's trust in state-run media among strategically important audiences, has created a situation where those audiences have unhindered access to the once marginal discourse. The latter is exerting a significant influence on those audiences' perceptions of reality, in fact, turning from marginal into the one that effectively opposes the official discourse. This situation is quite unprecedented for Russia.

Much of the recently published international academic literature on journalism, production, circulation and consumption of news, is devoted to digital hybrid ecosystems, such as Twitter and Facebook, which, according to most authors, now play the role of “gatekeepers” of news. Traditional ways of conveying to audiences what in the past was determined by professional journalists as important or worthy to be called “news” are no longer working. Modern algorithms applied by digital hybrid ecosystems allow users to determine what is presented to them and how it is presented. Moreover, after being published by professional journalists, news is often re-contextualised or re-interpreted by influencers who treat news content from different angles, with their own comments, criticisms, etc. Thus, the concepts of priming and setting a news agenda, as well as of controlling media discourse by a limited number of professionals are no longer relevant. The news diet of large segments of society is determined rather by the number of “likes” and views by other users, as well as by individual behaviour and preferences of a particular consumer.

Dr. Pierre-Emmanuel Thomann, professor in geopolitics, Lyon University 3 Jean Moulin and  ISSEP Lyon (France), and President of Eurocontinent (Belgium) submitted his paper on the theme “Artificial Intelligence, Digitalization and Global Geopolitical competition: What Role for Europe for a More Cooperative Agenda to Counter the Malicious Use of Artificial Intelligence”. 

The world is facing an increasing geopolitical fragmentation with the multiplication of actors, the reinforcement of power gaps between states and the changing of previous geopolitical hierarchies. Moreover, geopolitical confrontation is more and more the theatre of hybrid warfare including psychological warfare. In this context, digitalization associated with the emergence of AI used as a geopolitical weapon through the destabilization of international psychology security might contribute to determining the international order of the coming century, accelerating the dynamics of the previous cycle in which technology and power mutually reinforce each other. It will transform some paradigms of geopolitics through new relationships between territories, spatial-temporal dimensions and immateriality.

Not only the malicious use of AI at a tactical level by terrorist groups or states can have powerful effects in a conflict for geopolitical influence, but also more "neutral" aspects of AI powered digitalization will create new geopolitical hierarchies if it is used as a way that reinforces a monopoly by one or several states. Both have the potential to destabilize the system of international relations.

From the European point of view, it is admitted that the US and China will dominate AI and digitalization in the international geopolitical arena in the years to come. The main focus of the European Union regarding AI was so far on ethical and economic aspects and this is reflected in its main communication strategy. The EU commission has published in February 2020 a White Paper on AI and a Report on the safety and liability aspects of AI. The new documents do not address the development and use of AI for military purposes.

Regarding the risks concerning the malicious use of AI, the EU White Paper focuses mainly on the question of safety and liability of AI products that will circulate on the EU internal markets.

In 2019, France was also the first European state to publish a military AI strategy, as Defense was designated as a priority AI sector for industrial policy in the French 2018 national AI strategy. France’s approach to AI includes a strong geopolitical dimension as its wants France and Europe to avoid becoming dependant on the United States and China.

In Germany, the military, security and geopolitical elements of AI are not included in its 2018 National AI Strategy. However, the German approach on military AI focuses on arms control and debates disarmament.

This difference between French and German approaches is problematic for a strong common EU strategy as well as for common military industrial projects like the French-German new future Combat Air System fighter jet, supposed to include AI technology.

How can EU member states with all their diversity contribute to international cooperation to counter MUAI and protect IPS with other global actors like the US, China and Russia but also secondary actors? Maybe international cooperation based on inclusiveness, respect and reciprocity can be better achieved with a better geopolitical balance regarding AI and digitalization between global actors like the US, China, Russia and EU member states.

The lively discussion caused by the presented papers showed the novelty and importance of the problems of the psychological security caused by the implementation of artificial intelligence. The panel participants agreed to cooperate in joint events and publications.

- Kaleria Kramar MA (ISCPSC), Oleg Sarychev (ISCPSC)

(Version edited by ALAI)