Can machines think?
Less than a decade after decoding the Nazi encryption machine Enigma, and helping the Allied Forces win World War II, mathematician Alan Turing changed history a second time with a simple question: “Can machines think?”. Here we are, 70 years later, answering “yes” to that question. In fact, Artificial Intelligence is the branch of computer science that aims to replicate or simulate human intelligence in machines. Through a set of algorithms and learning processes, these machines became capable of formulating answers and performing actions that are based on the inputs received from the environment, as if they were human brains. Whether Turing could have been able or not to forecast such fast evolution, AI is a clear reality now, and it will inevitably impact tomorrow’s political and social scenarios. Before diving into AI machines’ strategic developments, it is important to explain how they work.
AI machines are mainly composed of a physical element (the Hardware), a digital element (the Software), and the huge amount of data they elaborate (the Big Data). Since AI needs large data processing capacities, the Hardware plays a very critical role. For this reason, modern companies have developed the so-called “supercomputers”, computers with accelerated pieces of hardware, capable of increasing the speed of the system that governs Artificial Intelligence. This allows to process a terrific amount of data, otherwise impossible with regular computers. The principle at the core of the technology is to gather as much information as possible in order to provide the best outcome.
These kinds of new technologies not only explain the future towards which our modern societies are heading, but they are already a concrete reality. Governments are investing huge amounts of money to develop innovative Artificial Intelligence; companies are already applying AI in many productive sectors, as well as in the public sphere, in the context of a wider digital revolution. Such important changes also affect the political sphere, and especially in terms of geopolitics, AI will be a constant driver in shaping the world’s future balance of power. Many applications of this technology to the public and political sphere are already in place, and we will go through some of them to illustrate the possible future developments and their consequences under different aspects.
Will machines be able to guarantee social control?
Although in Europe a social application of the Artificial Intelligence technology is still unreached, there are countries where it has been strongly implemented, and has already become a reality in citizens’ life. For example, China is the major global developer of AI, reflecting the President’s dream of “China 2.0”. Xi Jinping has clearly stated how Chinese future goals will be pursued through a general rise of technological standards, in order to make the country the major global provider of high techs. AI plays a crucial role in this regard, which is why Beijing has put such attention on its development.
One of the most popular applications of AI to social reality is facial recognition. It is already embedded in Chinese citizens’ life and it is exploited in many social fields. In big cities like Shanghai, every public aspect of the social life is controlled by intelligent cameras, from street security to work surveillance. As Giada Messetti states in her book “Nella testa del Dragone” (lit. “In the head of the Dragon”), the actual proportion between cameras and citizens is around 3/10, which can sound very impressive from a western point of view. Nevertheless, we must understand that Chinese people are both keen and used to these intrusive privacy measures, because they provide the citizens a safe and protected life environment. Moreover, these AIs also allow Chinese people to access consumptions more easily, which is an actual priority in Beijing’s modern society.
Even though Chinese people have a different view on privacy matters, there are some controversial uses of AI that can be traced back within the sphere of social control. A clear example is the facial recognition used as a way to guarantee “social security”. As there are no clear boundaries in the concept of social security, AI has been used also to recognize people considered keen to provoke “social instability”. Given such a vague concept, we must take into consideration that this technology can enhance social restrictions, such as censorship or social repression, especially in a single-party non-democratic country. Moreover, Beijing has always been clear about its policy of forcing every industry involved in the Chinese social life to collaborate with the government. This includes all the companies that provide the above-mentioned services.
AI technology is the key for a modern future, and it will probably play a key role for those countries that will be able to exploit its wide range of possibilities. Meanwhile, under every light there is always a shadow, and the use of facial recognition to oppress ethnical minorities, such as the Uighurs in Xingjian, is a clear example of the overwhelming power of the AI. Indeed, the constant tracking procedure – pursued through the collected data – allows the government to keep a constant eye on every aspect of people’s lives. Thus, there is no real freedom in what these minorities are able to do without the government knowing: facial recognition only paves the way to a deeper oppressive behaviour. The grey area around the exploitation of this new technology is related to its innovation in the global scene, but it is important to understand its great potential, and to set boundaries and rules to regulate AI’s implementation in social and public contexts.
Big Data’s strategic revolution
In our introduction, we understood how AI can exploit its wide possibilities thanks to a huge amount of data. Indeed, information and data are the so-called “oil of the future” due to their growing importance. AI is able to process enormous amounts of information that it receives, and to make the data both analysable and comprehensible. The speed of the AI’s calculation capacity allows it to produce valuable predictions about different issues. For example, it can lead to forecast the success of a new product thanks to the information about the market situation, or it can make previsions about more complex topics, such as economic macro-trends, thus letting decision makers act accordingly. Such a use of AI is the key to understand the influence that this technology will have on policymakers in the future.
The technological revolution of the internet pushed new horizons of tech developments, like 5G connection and smart cities. These innovations will be the gateway for the leap into the future of our societies, and data will be their common fuel. Thanks to AI, governments will be able to provide better services and to forecast future trends. Yet, in order to do so, they will necessitate to process the highest number of information possible. This allows us to understand that populous countries have an intrinsic advantage, as they can use their population as a huge pool to gather more data. In this way, they will have the possibility to develop more accurate forecasts and previsions about the future, turning the big data into a concrete political advantage on other competitors. The clearest example that we are already witnessing is China, which, being a world technology leader, will guarantee Beijing political advantages in forecasting future international trends. Thus, due to AI developments, big data supremacy will soon mean political supremacy as well.
Adapting Secret Service: how espionage will change due to AI
The activity of espionage is driven by states’ efforts to secretly collect information about foreign governments’ policies, decisions and intentions.
According to the US National Security Agency (NSA), intelligence works as a bridge between AI and espionage, dwelling upon the acquisition of precious information and its safeguard. It essentially comprehends the manner in which data is gathered, evaluated and dispersed. Intelligence can also be collected through technical means (TECHINT) such as imagery intelligence (IMINT), signals intelligence (SIGINT), measurement and signature intelligence (MASINT), and open source intelligence (OSINT). Nonetheless, with its development, the Intelligence services will boost their need for Artificial Intelligence and autonomous systems.
Firstly, the AI is of essential help to find the new upcoming threats, since it would be able to predict and alert in case of leakage of information. An example of how countries are getting ready to deal with the need of protecting their national data from being leaked is offered by the NATO’s Rapid Reaction Team (RRT). The RRT’s goal is to give assistance to NATO Member States or facilities that are subject to a cyberattack and to obtain an excellent responsiveness in real life whenever there is a data leakage. Jean-Francois Agneessens is one of the six members of the team, whose expertise varies from security audit to computer codes. According to him, ‘‘Cyber espionage, or dormant codes, which can disable national or NATO systems, pose new threats with a higher level of sophistication’’. In this scenario, the security team will need every possible help considering the complexity of exchanging security information.
The idea behind the creation of this team was given by the Alliance’s cyber defence policy in 2011, then improved by the 2014 Wales Summit. A cyberattack represents a threat to the harmony, growth and stability of the NATO Members, since it could trigger Article 5 of the Alliance treaty (the so-called collective defence clause). As a matter of fact, the same link between defence and AI was acknowledged by the European Parliament’s Committee of Foreign Affairs, which pushed the Member States to create a cyber-defence unit for a combined effort on common defence objectives. Indeed, according to the European Union Agency for Cybersecurity (ENISA), ‘‘cyber espionage is considered both a threat and a motive in the cybersecurity playbook’’.
Secondly, the AI is effective in “screening for anomalies in web traffic patterns or data”, potentially subjected to be attacked. Last but not least, experts highlight the importance of AI in machine translation and speech recognition, thus allowing analysts to move through various and different amounts of collected data. Nowadays, the technological evolution has even developed algorithms evolving just like organisms, thus being traceable for cyber espionage. An example is the web proxy debugging Fiddler, which helps users debug and analyse web traffic patterns.
These definitions highlight how cyber espionage is primarily based on operating through computer networks to obtain illegal and compromising data, usually in possession of the government or state’s institutions. Cyber espionage focuses on geopolitics through the acquisition of information in fields that are of primary importance for the State. For instance, in an economic espionage case, the US Department of Justice denounced Mr. Xiaoqing Zheng for spying General Electric, having stolen information in order to benefit from the improvement of two Chinese turbine companies. This was done by transferring the data into a USB, encrypting the information and sending them via mail.
Consequently, the General Electric case triggered large investments in Artificial Intelligence in order to prevent any other forms of espionage. A first example of the increased use of AI by the intelligence services concerns the US Open Source Enterprise, where the AI has taken the role that once belonged to human CIA’s readers and translators. Now machines are the ones collecting news from worldwide articles in order to direct geopolitical developments and potential crises in real-time.
Another interesting case of the AI social application is given by the Huawei biometric smart surveillance system in Belgrade. This system was created and presented in 2019, with thousands of cameras planted in the Serbian city, whose number has constantly increased up to this year. At the end of January 2021, the Serbian President Aleksandar Vucic suggested the introduction of AI into the courtrooms too, by collecting the case files and analysing them.
Anyway, the implementation of AI in Intelligence will need technological investments. For instance, deep learning is a technique based upon specific computer programs, neural networks, that go through huge amounts of data and remember the patterns they find along the way. Applications of this technique are oriented towards the development of self-driving vehicles. Seven out of ten companies have the intention of putting deep-learning-based AI as the base of their operations.
Therefore, it seems that espionage will never be the same, whether it will take the shape of biometric smart surveillance or language processing of evolving algorithms. Day by day, Artificial Intelligence is becoming the turning point that links the nations to the conquest of the new petrol oil: data.
AI on the battlefield: application to military technologies
‘‘The U.S. military operates more than 11.000 unmanned aerial systems and even more underwater, space, and terrestrial systems.’’
According to Foreign Affairs, “the U.S. cybersecurity units must deal with millions of bots on global networks as well as billions of Internet of Things devices”. These devices behave like sensors that, in order to operate, need their own diffusion system, and therefore, also their own intelligence. GEOINT (Geospatial Intelligence) Singularity is the proliferation of satellites and sensors that will make everything on Earth visible from above. Obviously, States will need to implement Artificial Intelligence in order to make all of this possible.
However, what mostly worries the analysts is not how autonomous a system may become, but whether or not it will operate in a comprehensible manner for humans. Indeed, as neuroscientists believe, machines learn to emulate how the brain works when receiving new information, and rearrange their digital innards as a feedback to patterns they have spotted in the data. As a result, the machine designer cannot know, once the machine is ready, what will actually be the outcome. This becomes extremely crucial in warfare, for instance. In fact, though artificial intelligence may incur in some mistakes, “there is very little that a human can do to understand how the system [made] its decisions”.
To summarize, the Internet of Things, sensors and proliferation systems are the new frontier of the military technologies. However, humans should retain control over them or, at least, better understand the AI’s decision-making process.
A chess match where the victory is the AI leadership on an international level. This is the competition between the two great powers: the USA and China. On one hand, there is the USA with Google’s DeepMind, on the other, China with Tencent. Machine learning is the battlefield where ideas meet challenges. Two are the main obstacles: adoption of the technology and talent pool for development.
Google’s DeepMind is an Artificial Intelligence agency that works on the Internet of Things field with human-like intelligence in order to achieve developing computers and systems that process information as humans do. Initially it started as an AI machine able to challenge gaming world champions. According to Kai-Fu Lee, American tech giants “tried and failed to win the Chinese market” because they treated the Chinese competition like any other. This is visible with the Chinese tech company Tencent, which focuses on training and developing the talent of the workers by themselves through various projects in order to attract talent. According to the Wilson Center, “the U.S leadership in the Artificial Intelligence research and development may be challenged by China’s policy planning”, focused on swiftly improving AI techniques. As a matter of fact, the Chinese government is increasingly supporting the AI implementation both on the domestic and on the international spheres.
On the occasion of the China Development Forum held in 2017, the McKinsey Global Institute published a report which analyses the AI’s ability to feed China’s productivity growth and the country’s workforce. For China, the potential of developing AI stands in not having to reprogramme the computer systems and giving strategy for learning, allowing new data inputs to be adapted without further programmatic efforts. AI and machine-learning systems are nowadays present in various fields, from finance to healthcare, making China one of the most important global hubs for AI development.
The large population and various industry mix can create huge volumes of data and arrange an enormous market, thus Chinese largest companies are the ones investing into AI research and development. The boost that these technologies give is essential for the stability of China’s economic growth and population. The crucial points stand into the integration of AI techniques into the more traditional industry sectors. In order to obtain this result, there will be a need for strategic awareness, implementing technical skills and breaking the barrier of additional costs. According to the McKinsey Global Institute, China’s strategy sets upon the following steps: “building a robust data ecosystem, spurring adoption of AI within traditional industries, strengthening the pipeline of specialized AI talent, ensuring that education and training systems are up to the challenge, and establishing an ethical and legal consensus among Chinese citizens and in the global community”.
If, on the one hand, AI can improve several fields, going from environment to education, on the other hand, the legal and safety matters need to be targeted with particular attention. Thus, China has both the qualities and the potential to create an international coordination in the research and development of AI, assuring a contribution to global growth and human welfare.