Artificial intelligence and international psychological security:
academic discussion in Khanty-Mansiysk and Moscow
The development of artificial intelligence and machine learning, embedded systems and devices, the Internet of things, augmented and virtual reality, big data analysis (data science) and cloud computing, block chain, etc. stimulate the transition to a new technological order. At the same time, positive expectations associated with scientific and technological progress are combined with clearly perceived threats of the approaching future, which are described and analyzed by representatives of various discourses – from mass media to academic and political circles. The development of artificial intelligence (AI) is a subject of discussion in the broad context of economic, political and related social transformations. The importance of the introduction of AI is now recognized by almost all states and international organizations. However, new threats in the field of psychological security, caused primarily by global international tensions and the growing influence of non-state actors, including criminal, terrorist organizations, are closely linked with the use of new tools provided by AI, in the field of communication in particular.
The attention of the academic community to this range of problems is evidenced by the active discussion that took place at the panel discussion “Malicious use of artificial intelligence and international psychological security” of the UNESCO Conference in Khanty-Mansiysk and continued at the eponymous research seminar at the Diplomatic Academy of the Ministry of Foreign Affairs of the Russian Federation.
The II International Conference “Tangible and Intangible Impact of Information and Communication in the Digital Age” was held within the framework of the UNESCO Intergovernmental Information for All Programme (IFAP) and XI International IT Forum with the participation of BRICS and SCO countries in Khanty-Mansiysk on June 9-12, 2019. The conference was organized by the Government of the Khanty-Mansiysk Autonomous Okrug – Ugra, Commission of the Russian Federation for UNESCO, UNESCO Intergovernmental Information for All Programme (IFAP), UNESCO Institute for Information Technologies in Education (IITE), Russian Committee of the UNESCO IFAP, Diplomatic Academy of the Ministry of Foreign Affairs of the Russian Federation and the Interregional Library Cooperation Centre.
A number of academic institutions provided academic support for the event. These are the International Center for Social and Political Studies and Consulting (ICSPSC), the European-Russian Communication Management Network (EU-RU-CM Network) and the Russian – Latin American Strategic Studies Association (RLASSA).
The conference was supported by the publishing house “Ugra News”, the Institute for Political, Social and Economic Studies – EURISPES (Rome), the Association of Studies, Research and Internationalization in Eurasia and Africa – ASRIE (Rome), the Geopolitics of the East Association (Bucharest), the International Association “Eurocontinent” (Brussels) and the International Institute for Scientific Research – IIRS (Marrakech).
The Governor of Ugra Natalia Komarova at the opening the panel discussion “Malicious use of artificial intelligence and international information and psychological security”
The Government of Ugra attaches great importance to the development of cooperation with UNESCO, which has been counting down since 2001. The Governor of Ugra Natalia KOMAROVA took part in the conference. Opening the panel discussion “Malicious use of artificial intelligence and international information and psychological security”, the head of the region recalled that a year ago the Ugra Declaration adopted at the first UNESCO conference included proposals for the preparation of a world report on socio-cultural transformations in the digital age, the formation of educational programs relating to ethical, legal, cultural, social aspects of life. “In the modern world, artificial intelligence is rapidly gaining popularity, the issues of its implementation for good purposes are discussed, concerns about the negative impact are expressed, – Natalia Komarova continued. – Russian President Vladimir Putin, speaking at the recent Economic Forum in St. Petersburg, said that by 2024 the world market of products using artificial intelligence will grow by almost 17 times”. She stressed that in the era of global transformations, competition for resources, and especially for human ones (intelligence), is growing.
According to the Governor, in connection with the mass involvement of people in the global communication space, the subtlety of the ideological setting of society is of particular importance: “Social identity is forced to comprehend changes, accepting, mastering changing values. Experts note that in a situation of instability of the value landscape identity can not be stable – it becomes fluid. In this case, the English sociologist Anthony Giddens notes, identity is constructed, modernity is distinguished by ‘appetite for new’, and this new encourages us to act, including the development of artificial intelligence, which in turn requires the creation, use of information security, prevention of cyber attacks, protection against data leakage. At the same time, according to Natalia Komarova, there is a need to ensure psychological security.
Vice-Chair of the Intergovernmental Council for the UNESCO Information for All Programme (IFAP); Chair of UNESCO IFAP Working Group for Multilingualism in Cyberspace; Chair of the Russian IFAP Committee, and President of the Interregional Library Cooperation Centre Evgeny KUZMIN recalled that since 2001, Ugra has hosted major information for all conferences. “This is a serious contribution of the Khanty-Mansiysk Autonomous Okrug and the whole of Russia to the implementation of this flagship intergovernmental program of UNESCO. The events were devoted to the discussion of such important issues as the preservation of the languages of the peoples of the world and the development of linguistic diversity in cyberspace, the formation of open government, transparency of governance, the interaction of governments and the population by improving information literacy among officials and citizens,” he stressed.
The Interregional Library Cooperation Centre under the leadership of Evgeny Kuzmin made significant efforts to ensure that the conference was held at a high academic level. In total, the conference was attended by representatives of 35 countries from all over the world.
From left to right: Marco Ricceri, Evgeny Pashentsev, and Dorothy Gordon
With academic support of EU-RU-CM Network the conference was attended by coordinators and network members: Darya Bazarkina (Russia), Evgeny Pashentsev (Russia), Olga Polunina (Russia), Marco Ricceri (Italy), Gregory Simons (Latvia/New Zealand/Sweden), Pierre-Emmanuel Thomann (Belgium) and Marius Vacarelu (Romania).
At the opening of the conference Evgeny PASHENTSEV, Leading Researcher, Diplomatic Academy of the Ministry of Foreign Affairs of the Russian Federation; Director, International Centre for Social and Political Studies and Consulting (Moscow, Russia), Coordinator of the European-Russian Communication Management Network (EU-RU-CM Network), senior researcher, St. Petersburg State University presented a paper on “Artificial Intelligence: Current and Promising Threats to International Psychological Security”. We further provide the full text of his paper.
International security today is under threat due to destructive processes in the economic, social, military and other spheres of public life. Negative processes are developing at the national, regional and global levels. It is essential to have an adequate understanding of the existing problems by all sectors of society. But this goal is far from being realized. Perhaps because of the vested interests and/ or irresponsibility of the ruling elites who seek to hide the truth about the real state of things. Or, much worse, because of their intellectual and moral degradation, the elites are unable to respond to the increasingly obvious threats through a system of effective action, not even for the sake of social progress, but for the sake of their physical self-preservation.
The low level of civil self-organization of society, the lack of established progressive counter- elites testifies to the crisis not only of the "top", but also of the "bottom" strata of the society, i.e. of the civilizational crisis. The only way to solve the problems facing humanity is to have access to information, to the modern possibilities of its processing, analysis and dissemination. On this basis it is possible to offer scientifically-proved models of progressive development of mankind, their discussion, both in the professional environment, and at public discussion forums. Artificial intelligence can be of great help in the processing, analysis, verification of research results and implementation of relevant social development programs. Human beings’ intellectual creative abilities, will, aspirations and feelings "Weak AI" will not replace, and "Strong AI" equal or superior to the modern human mind – is a matter of the future. Unfortunately, the rapidly growing practice of using AI to manipulate public consciousness at the international level, once again testifies to the large and dangerous potential of the negative consequences of the use of new technologies.
International psychological security (IPS) means protecting the system of international relations from negative information and psychological influences associated with various factors of international development. The latter include targeted efforts by various state, non-state and supranational actors to achieve partial/complete, local/global, short-term/long-term, and latent/open destabilization of the international situation in order to gain competitive advantages, even through the physical elimination of the enemy.
International actors engaging in hybrid warfare exert negative direct and indirect impacts on the enemy’s public consciousness and, often, on themselves, their allies and neutral actors. For example, economic sanctions are intended not only to financially weaken/destroy the enemy, but also to reduce the readiness of target groups for further resistance by increasing the enemy’s economic problems. Military-political confrontation with the enemy, based on aggressive interests and mass genocide against other nations, causes irrecoverable damage to the mentality and psyche of the aggressor country’s population. At the same time, psychological warfare (PW) is always aimed at delivering direct (although often latent) blows to the enemy’s public consciousness and achieving (through victory in this sphere) a bloodless total victory over the enemy. In fact, the modern globalworld is witnessing hybrid warfare in the system of international relations, which has never completely stopped throughout history; rather it has had natural periods of exacerbation. We have clearly entered a long-term transition period in the development of humanity and the system of international relations in particular, which is accompanied by irregularly growing PW.
In our opinion MUAI can allow hostile actors to be more successful than so far in:
• provoking a public reaction to a non-existent factor of social development in the interests of the customer of psychological impact. The target audience sees something that doesn't really exist.
• presenting a false interpretation of the existing factor of social development and thus provoking the desired target reaction. The audience sees what exists, but in a false light.
• significantly and dangerously strengthening (weakening) public reaction to the real factor of social development. The audience sees what exists but reacts inadequately.
We can suggest the following MUAI classification according to the degree of implementability:
• current MUAI practices;
• existing MUAI capabilities that have not been used in practice yet (this probability is associated with a wide range of new rapidly developing AI capabilities — not all of them are immediately included in the range of implemented MUAI capabilities);
• future MUAI capabilities based on current developments and future research (assessments should be given for the short, medium and long term);
• unidentified risks, also known as “the unknown in the unknown.” Not all AI developments can be accurately assessed. Readiness to meet unexpected hidden risks is crucial.
It is important and necessary to use independent teams of different specialists and AI systems to assess MUAI capabilities.
We can also propose the following MUAI classifications:
• by territorial coverage: local, regional, global;
• by the degree of damage: insignificant, significant, major, catastrophic;
• by the speed of propagation: slow, fast, rapid;
• by the form of propagation: open, hidden. Among the possible threats of MUAI (see moredetailed: Bazarkina and Pashentsev, 2019), which can cause a serious destabilizing impact on the socio-political development of a country and the system of international relations, including the sphere of IIB, are the following:
•The growth of integrated, all-encompassing systems with active or leading AI use increases the risk of malicious takeover of such systems. Numerous infrastructure facilities, for example, robotic self-learning transport systems with AI-based centralized management, can be convenient targets for high-tech terrorist attacks. If terrorists seize control over the transport management system of a large city, this may lead to numerous casualties, cause panic and create a psychological climate that will facilitate further hostile actions.
•The reorientation of commercial AI systems. Commercial systems can be used in harmful and unintended ways, such as deploying drones or autonomous vehicles to deliver explosives and cause crashes (Brundage, et al., 2018, p. 27). A series of serious disasters, especially those involving celebrities, may cause international media hype and damage IPS.
•Attacks further removed in time and space. Physical attacks are further removed from the actor initiating the attack as a result of autonomous operation using AI (Brundage, et al., 2018, p. 28). The surprise effect of such attacks may destabilize the system of international relations. For example, nuclear devices can be simultaneously set off from afar in different countries of the world without direct human participation. Officials of all countries that possess modern technologies speak of the need to retain control over the combat uses of AI systems. This is understandable, since no government, reactionary or progressive, wants to lose control over their weapons. But this does not apply to non-state actors: for example, a group of techno-religious maniacs who want to eliminate humanity will have an increasing chance of success due to the continuous improvement of AI, the creation of complex cross-border AI systems, the propagation of new technologies, and other factors.
•The creation of ‘deepfakes’. ‘Deepfake’ (a portmanteau of “deep learning” and “fake”) is an AI-based human image/voice synthesis technique. Many celebrities, including Scarlett Johansson, Maisie Williams, Taylor Swift and Mila Kunis, have fallen victim to deepfake pornography. Deepfakes hobbyists have begun using this technology to create digitally-altered videos of world leaders, including U.S. President Donald Trump, Russian President Vladimir Putin, former U.S. president Barack Obama and former presidential candidate Hillary Clinton. Experts warn that the videos could be realistic enough to manipulate future elections and global politics as early as 2020 (Palmer, 2018). Deepfakes can be used in psychological warfare to provoke financial panic and trade or hot wars. Fake videos of Israeli Prime Minister Benjamin Netanyahu or other government officials – for instance, talking about impending plans to take over Jerusalem’s Temple Mount and Al-Aqsa Mosque – could spread like wildfire (The Times of Israel, 2018). Just as dangerous is the possibility that deepfake technology spreads to the point that people are unwilling to trust video or audio evidence (Waddel, 2018).
•‘Fake People’ technology. After the sale of the first AI-generated painting in early 2018, deep learning algorithms now generate portraits of non-existent people. The NVIDIA company has recently published the results of the work of a generative adversarial network (GAN) trained to generate images of people (Karras, Laine and Aila, 2018). The technique is based on an infinite collection of images of real faces; this is why a neural network recognizes and applies many fine details in its work. It can generate hundreds of faces with glasses, but with different hairstyles, skin textures, wrinkles and scars, and add age signs, cultural and ethnic features, emotions, moods or effects of external factors, such as wind in the hair or an uneven tan. Back in 2017, NVIDIA experts held a similar experiment, but the images of faces they got then were blurry and were easily recognized as fakes. Today, neural networks are incomparably better and generate faces in high resolution. They can easily produce, for example, an image of a non-existent illegitimate child of a celebrity, with a perfect family resemblance, as a provocation.
•Agenda setting and amplification. Studies indicate that bots made up over 50 percent of all online traffic in 2016. Entities that artificially promote content can manipulate the “agenda setting” principle, which dictates that the more often people see certain content, the more they think it is important (Horowitz, et al., 2018, pp. 5-6). Reputational damage done by bots during political campaigns, for example, can be used by terrorist groups to attract new supporters or organize assassinations of politicians.
•Sentiment analysis is a class of content-analysis methods used in computational linguistics to identify emotionally loaded words in texts that reveal the author’s opinion of the topic. Sentiment analysis is done on the basis of a wide range of sources, such as blogs, articles, forums, polls, etc. This can be a very effective tool in PW.
The development of Artificial Emotional Intelligence (AEI). Here research is conducted not only in the field of artificial, but also natural, human intelligence. The research here develops in several areas: firstly, the recognition of emotions of humans and animals, secondly, is the analysis, interpretation of these emotions and the necessary for that techniques. For this purpose, machine learning and big data analysis are used. The third direction is the reproduction of emotions in robotic systems. If we talk about the reproducibility of emotions in robotic systems, the Japan has achieved great success in this area. This direction is well applicable for example to the field of care for the elderly. The full creation of emotional AI is possible only within the framework of the creation of "Strong AI". Unfortunately, the development of emotional AI within the framework of "Weak AI" also poses many threats to the IPS as it opens up new forms of control over the human mind through AI, including the provocation of mass riots.
•AI, machine learning and sentiment analysis make it possible to predict the future by analyzing the past — quite a holy grail for the financial sector or government planning agencies. But various state and non-state actors can potentially use this possibility for MUAI. Particularly important are prognostic weapons: predictive analytics methods based on big data and AI, which make it possible to correct the future from the present in one’s own interests and contrary to the objective interests of the target. For example, the Intelligence Advanced Research Project Activity (IARPA) launched the Early Model Based Event Recognition Using Surrogates (EMBERS) program in 2012 to forecast socially significant population-level events, such as incidents of civil unrest, disease outbreaks, and election outcomes. For civil unrest, EMBERS produces detailed forecasts about future events, including the date, location, type of event, and protesting population, along with any uncertainties. The system processes a range of data, from open-source media, such as Twitter, to higher-quality sources, such as economic indicators, processing about five million messages a day. The system delivers more than 50 predictions about civil unrest alone for 30 days ahead (see: Doyle, et al., 2014).
•Growing threats from phishing. AI allows to increase dramatically data processing speed and respond faster to people's expectations, which makes phishing more dangerous. Progress in automated spear phishing has demonstrated that automatically generated text can be effective at fooling humans, and indeed very simple approaches can be convincing to humans, especially when the text pertains to certain topics such as entertainment (Brundage, et al., 2018, p. 3, 46). Main methods of using artificial intelligence hackers are phishing, spear phishing and whaling, i.e. phishing focused on senior managers responsible for financial decision-making. •Computer games using AI can also increase the effectiveness of psychological impact, especially on children and adolescents. AI is already actively used in the creation of computer games. That computer games can have a certain manipulative affect, it is known well for a long time, however the analysis of use of AI for these purposes – one of perspective tasks of researchers. From the point of view of IPS, special attention should be paid to computer games, which are widely distributed in many countries of the world.
•It can be imagined that due to a combination of psychological influence techniques, sophisticated AI systems and big data, synthetic information products could emerge in the near future that would be similar in nature to modular malicious software. However, they will have an effect not on inanimate objects, social media, etc., but on humans (individuals and masses) as psychological and biophysical beings. These synthetic information products will contain software modules that will drive large numbers of people into depression. After that, suggestive programs will latently come into action. Appealing to habits, stereotypes, and even psychophysiology, they will encourage people to perform strictly defined actions (Larina and Ovchinskiy, 2018, pp. 126-127).
However, any of the above mentioned threats can also be more effectively neutralized with a help of AI. For example, Swisscom Innovations developed and trained an artificial intelligence based phishing detection system. It predicts reliably whether a formerly unknown website contains phishing or not (Bürgi, 2016). Another programme, Lookout Phishing AI continuously scans the Internet looking for malicious websites. Lookout Phishing AI detects the early signals of phishing, protects end users from visiting such sites as they come up, and alerts the targeted organizations (Richards, J., 2019).
The task today is to repel threats from the real and constantly developing “weak” artificial intelligence, which is a threat not in itself, but because of the actions of antisocial external and internal actors that turn it into a threat to the international security. In the not so distant future, there may be problems associated with “strong intelligence”, the possibility of which in the coming decades, forecast more and more researchers.
Blinnikova, N., 2018. Emocional'nyj II podskazhet, kogda na rabote luchshe pojti popit' chaj, i pomozhet borot'sja so stressom (Emotional AI will tell you when to go to work to drink tea, and will help fight stress). ITMO News.
Terrorist propaganda adapts to the expectations of the target audience, of which potential and actual recruits are an important part. Finally yet importantly, to recruit young people into its ranks, IS publishes materials aimed at developing a more “high-tech” image of a terrorist, which combines the features of a fanatic and, for example, a skilled hacker. These phenomena could be a prologue to a new, much more dangerous phase of terrorist activity, in which terrorist acts could become much more destructive and their perpetrators, operating at a distance with the advanced technology, would become extremely difficult to detect. In this regard, it is worth mentioning the magazine “Kybernetiq”.
It is advisable to widely use predictive analytics mechanisms by state and supranational agencies to prevent social unrest (through timely social, economic and political measures to achieve social stability in the long term). Among measures not directly related to AI (but potentially optimized with its help), governments should develop long-term policies towards the social integration of people of different religions and sects into socially significant projects. In countries and regions experiencing social and economic instability, apart from taking measures to improve the well-being of citizens, governments should explain to the population the economic and political goals of terrorist organizations and the essence of the ideology of terrorism. The larger the scope of terrorist activities, the higher international agencies should predict and take such measures.
Fatima ROUMATE, Associate Professor, Mohamed V University; President, Institut International de la Recherche Scientifique (Marrakech, Morocco)
Malicious Use of Artificial Intelligence: New Challenges for International Relations and International Psychological Security
Nowadays, AI offers new opportunities for international and bilateral cooperation, and facilitates the inclusion of all actors within global governance. However, the malicious use of AI represents a threat to the international psychological security whether we are speaking about social, economic or military activities. In fact, this threat is an important feature of the new cold war characterized by the race toward AI. A new international order is in a progress considering the rise of new technological and economic forces which means emerging of new players and new rules of the international relations.
We’ll have a closer look at the impact of the malicious use of AI on the international community and the challenges the international community faces today considering the AI race in economic and military areas.
International actors are using AI to achieve their specific beneficial goals. However, they are investing more efforts to limit their vulnerabilities. The consequence is that international society faces the psychological impact of non-trusted information which influences policy-maker decision and political changes in global affairs.
Malicious use of AI leads us to think about one of the most important negative impacts, especially the attacks on democracy. In fact, AI is not only expanding existing threats, but it’s creating new threats. Spear phishing attacks, for example, increased significantly since 2016 in several countries such as Canada, France, Italy and the USA where attacks against specific targets accounts for more than 86% of all phishing attacks.
There is a new voice technology that can reproduce a believable fake voice. There is also the concept of machine learning software that creates fake videos. Moreover, “AI systems are expanding the phishing attacks space from email to other communication domains, such as phone calls and video conferencing”. This shows how it’s challenging for diplomats and countries to check the trustworthiness of the information and its sources before making a decision.
Governments invest in malicious use of AI, first, for surveillance and defense, second, by other states and non-states actors to create or support social movement aimed at specific political changes. In systems that combine data from satellite imagery, facial recognition-powered cameras, and cell phone location information, among other things, AI can provide a detailed picture of individuals’ movements as well as predicting future movement and location. It could therefore easily be used by governments to facilitate more precise restriction of the freedom of movement, at both the individual and group level and by foreign actors who are targeting political changes. Voting behavior and election campaigns are also influenced through social media.
Malicious use of AI can influence a lot of domains as defense, diplomacy, cyber security, economic and financial sector. According to the 2017 Official Annual Cybercrime Report, cyber crime costed $3 trillion in 2015 and it’s estimated that cybercrime will cost $6 trillion annually by 2021.
The malicious use of AI creates new challenges for states as an original actor in international relations. This invites researchers and policymakers to rethink many concepts linked to the state’s notion as sovereignty, diplomacy and security, considering the appearance of new notions as artificial diplomacy, cyber security, cyber war...
AI is creating big changes in international relations. It facilitates the integration of new actors in global issues, especially in this age characterized by the diffusion of power in international society and the expansion of transnational relations. In fact, in the age of AI, it is necessary to rethink all institutions outside and inside countries. The massive interconnection between all actors imposes the need to update diplomatic tools.
As for the influence of AI systems in global affairs, Hillary Clinton argues that the use of information and communication technologies has an influence on global debates and is playing greater roles in international affairs for good and bad.
In this sense, the future of international psychological security is conditioned by the State’s response to the challenges imposed by the cyber era. For that, the American security strategy focuses on the improvement of strategic planning and intelligence. At that juncture, real coordination between States and transnational corporations specialized in ICT (GAFA in USA and BAT in China) is a sine qua non-condition considering advances in artificial intelligence which are reshaping the practice of diplomacy.
The principal goal of the competition between China and USA is racing toward technological sovereignty which means, according to Nicholas Westcott, having a seat at the international table at the age of AI.
Malicious use of AI imposes new challenges related to international law and human rights, especially with the charter of principals and human rights in the internet which recognizes the access to the internet as a fundamental right. AI age is a new phase in the development of international law which becomes heavily traditional. In the same context, the appearance of the Lethal Autonomous Weapons (LAWs) creates a controversial discussion between States and it requires an urgent review of the use of force as it was cited in the UN charter. States competition toward LAWs lead us to think that current trade crisis between China and USA can be escalated to an open military conflict with the use of AI weapons.
First, the future of humanity will be decided by no state actors when they will own LAWs. Second, all these new technologies are growing faster than international law and diplomacy. Thus, international law norms such as those concerning the use of force and defense need to be revised.
States need to invest more on the LAWs to prevent violation of international humanitarian law. According to the Report of the Special “Rapporteur” on extrajudicial, summary or arbitrary executions, States must be transparent about the development, acquisition and use of armed drones. The goal is to ensure international psychological security which is a sine qua non-condition of international security.
The growing investment in AI for commercial and military will expand the challenges and threats to international psychological security. These challenges are significant because AI is growing rapidly while the development and updating of international mechanisms is very slow. This leads us to another challenge which is the creation of right balance, first between commercial and military funding dedicated to AI and second between investment in AI and protection of human rights in peace and in war.
Malicious use of AI invites all actors (States, international institutions, NGOs, transnational corporations and individuals) to collaborate and give a written riposte in the political, juridical and institutional level. The goal is to ensure international psychological security.
The challenges imposed by malicious use of AI are pushing international society towards a new global order with several and fundamental changes of players and rules in the international game.
Frederic LABARRE, analyst and Education management consultant at the Royal Military College of Canada, co-chair of the Regional Stability in the South Caucasus Study Group (Partnership for Peace Consortium) (Canada)
The Mechanics of Social Media and AI-aided Radicalization: Impact on Human Psychology (A digest from "Mapping Social-Media Enabled Radicalization: A Research Note" by P. Jolicoeur and F. Labarre (2017). The paper was sent to and presented at the research seminar at the Diplomatic Academy in Moscow, June 14th, 2019).
In 1957, at a moment when the discipline of political science was being developed, famed scholar David Easton came up with a systems analysis approach to examine and explain political action. At the time, his effort was ground-breaking, and for a while, hindsight looked on the systems approach of political analysis as hopelessly descriptive, and not analytical at all. Critics said that, for one, a systems approach could only explain the functioning of democratic systems. Sixty years later, with most regimes on the planet being democratic, it would seem that Easton’s approach is more relevant than ever. All the more so since half of the individuals on the planet, thanks to technology, carries this democratic power in their pocket; through their cell phones and devices.
As we have argued above, technology expands the individuals’ horizons and seemingly provides direct and instantaneous access to the political system. The neat compartmentalization that existed before the advent of the internet is no longer possible. Everything is instantaneous, and everyone has a voice. Clearly, this puts added pressure on elected officials. However, it also provides them with tools to channel demands. Contrary to what the Internet promised some quarter of a century ago, individuals are not exposed to many facets of a story or event, enabling them to make more “optimal” choices. Today, technology produces a perversion of democracy at the individual level; electronic systems deliver what the individual wants as opposed to what they need.
This is because individuals are their own systems. They are biological systems. Biological systems are subject to demands and pressures, and, like any other system, produce outputs and decisions which inform future demands from their immediate environment.
Groups that are adept at seizing upon this verity also leverage the power of technology to further pressure individuals, so that the individual is, wittingly or not, “enrolled” into a project or a political vision which is not their own initially.
Individuals, themselves bombarded by demands of all sorts, instinctively seek to make sense of the world by seeking reassuring biases. They will be less likely to challenge their own views. They will tend to keep company with like-minded people in communities of thought. These communities, however, are not always physical; they are often remote, brought perceptually closer thanks to the ubiquitous Internet and social media.
The problem comes when algorythms begins “feeding” individuals with “expected” support, thereby reinforcing pre-existing biases within individuals. Social-media provides a wealth of information on individual habits, allowing virtual communities to provide individuals with messages and images that are soothing and apparently give meaning and structure to what is seemingly a raw and chaotic world. The recent scandal involving Cambridge Analytica’s role in pushing negative messaging on targeted audiences is a case in point. The current malaise with fake news has not lifted the veil from peoples’ eyes; they have merely reinforced perceptual biases; it is the others’ news outlet which are fake, never our own.
Under pressure, social media like Facebook have resorted to artificial intelligence (AI) to reduce the incidence of fake, hateful, or radical content on its site. But AI, currently in the form of Bots and other automatic trolling devices, can also be leveraged to produce and reproduce radical content exponentially. This seems to be a losing battle, and the human mind never feels overwhelmed, as the content is always so emotionally fitting. Very soon, it will be difficult to determine what is genuine content from what is human-generated content.
Be that as it may, the output (in Eastonian terms) – the radical decisions – will always be human, and so radicalization is generated from the following path of exposure.
1) The person interacts “normally” in life and online. Choices made there are reflected in search histories, websites and other visits.
2) Data generated finds itself in data aggregators (i.e. contribute to “Big Data” pools).
3) Data aggregators inform radical groups of communities and individuals’ vulnerabilities and psychological predispositions. These “markers” may be social, economic, political and ethnic.
4) Physical and electronic messaging (by community and local political leaders, but by the media as well) begins by reflecting existing beliefs, reinforcing them.
5) Trolls and bots intervene in social media to activate (some old hands in Russia might say “agitate”) individuals. Some trolling might be favourable, other might be unfavourable to an idea. What matters is generating belief and feeling among the target population
6) AI takes over by self-generating messaging on social media, and also by directing and redirecting advertisement, searches, etc. towards “expected” outcomes. Individuals become progressively isolated within the pool of opinion. The outcome of this is that the people start to believe that there is only one stream of dominant opinion, or that a minority problem is ever present.
7) Psychological radicalization is achieved. Within a certain period of time, the individual takes steps that bridge the gap between the consumption of radical messaging and acting out a program propagated by ideologues.
The steps outlined above are the same for any program of “sale” of any idea. Whether it is to buy a car or medication, or changing the world through voting for a candidate, going down in the street, or planting a bomb in a café. Anyone and everyone is vulnerable in the same way. Only the socio-political outcome is different. And our definitions of the act. But that is another story.
Without time to reflect, without reasoned contact with competing or contrary opinion, and yet, even with assurances of perfectly clean data and statistics on a problem, individuals will always side with their preferred biases. The aim of the state is to avoid unnecessary bloodshed or upheavals. But technology provides other states and groups with the power to cause mayhem elsewhere. Technological advances in communications are not merely a double-edged sword; it’s only a blade with no handle, sure to slip from the bloody hand that wields it.
Aleksandr RAIKOV, Leading Researcher, Institute of Control Sciences, Russian Academy of Sciences (Moscow, Russia)
Strong Artificial Intelligence, Its Features and Ethical Principles of Safe Development
Artificial Intelligence (AI) is currently developing in a digital economy. It increasingly penetrates the socio-humanitarian and industrial sphere, helps to resolve a state and municipal government’s issues.
AI is a technology that enhances a person’s creative possibilities and helps him in his work. AI makes it possible to understand and use the power of the human mind, to get closer to the mystery of the human spirit. However, before the AI was a harmless human’s helper in routine, now it has already become a dangerous competitor for any employee.
However, the AI capabilities are expanding and deepening. It infiltrates deeper into the secrets of the sensual and emotional human sense levels, human’s meditative abilities, as well as the collective unconscious. And with this, features of the next generation of AI - Artificial Super-Intellect (ASI) begin to appear: “Intellect, which is much smarter than the best human mind in almost all areas, including scientific creativity, wisdom and social skills”. With the advent of ASI, its danger to society is not excluded. And on the way of its creation there can meet traps, hit in which is capable of causing irreparable damage to society.
The first trap is the digitalization trap, when a continuous (analog) image of reality in a computer is replaced by a digit (bits and bytes). And there is no matter how accurately a computer restores a continuous signal to its individual points, the error of computer models accumulates. The digital signal has a limited frequency spectrum. Because of this in particular a number of tasks on supercomputers are solved in months, instead of fractions of a second. And more importantly, such a signal is unable to reflect the full depth of human emotions and feelings, which ultimately can lead to a decrease in the level of culture and spirituality of society.
The second trap is the rationality trap. A human tries to find a rational grain everywhere: at home, at work. He analyzes. Analysis is the division of the whole into the parts. Synthesis is a reverse operation. This is a much more complicated operation that requires the connection of a human’s creative abilities. With a purely rational (algorithmic) approach to the synthesis, the possibilities of obtaining good solutions using AI are very limited. As a result, the risk of growing errors in solving vital problems, especially strategic ones, increases.
The third trap is the causality trap. Modern AI uses logic and statistics in its conclusions, which reflects causal connections and parameters correlations. The classical science canons require this. Forecasts are often made on the basis of experience accumulated over a certain period of time and established trends in the development of events. But life often offers to solve unexpected problems and in completely new circumstances, it behaves in a different illogical way, which modern AI cannot realize.
The fourth trap is phenomenological. Natural human intelligence is enhanced by emotions and feelings. There are also deeper levels of consciousness – meditative and transcendental. An even more complex phenomenon is the collective unconscious. These phenomena are sources of mysterious insight - instant comprehension of the whole, afflatus and insight of the human mind.
They are characterized by complete nonformalizability, illogicality, intensity, duration, objectivity, tonality, etc. Traditional AI cannot yet embrace these levels of consciousness.
The listed traps (the list is incomplete) are due to the stereotypes of conducting scientific research, insufficient coverage of disciplines and the lack of relevant international collaborations. But sooner or later, these limitations in the development of AI will be removed and the ASI will enter the arena. This requires the development of interdisciplinary basic research, a more critical attitude to digitalization, immersion of information models in infinite-dimensional spaces, removal of contradictions between quantum mechanics and the theory of relativity, appeal to the potential of Space and much more. It is also necessary to start teaching people the future!
The possible dangers associated with the future appearance of ASI make you think about their warning. Scientists, government and public figures, professionals and experts develop ethical principles that should be followed in order for the development of AI to move in a safe and moral manner. For this, in particular, the well-known Asilomar AI Principles have already been formed.
To support these principles, as well as taking into account the public and state significance of the issue of possible risks of ASI development, which may increase in unpredictable and abrupt manner, we consider it reational to offer the government authorities, the scientific and expert- analytical community the following minimum set of principles in the development of ASI:
1. The most important goal of ASI development should not only be to increase the efficiency of labor, but also to master the deep