“Stephen Hawking, the renowned British physicist, brilliantly summarized the main problems associated with AI: ‘Whereas the short-term impact of AI depends on who controls it, the long-term impact depends on whether it can be controlled at all.'”
he emergence of artificial intelligence (AI) will mark a turning point in the history of the human species. AI will not only lead to improvements in our daily lives and the way we interact with others; it is going to bring about dramatic societal upheavals—and force us finally to define our nature and determine what we are living for.
Most narratives about AI and politics are centered around election hacking, data surveillance and privacy issues, yet this focus disregards considering the new opportunities presented by AI. Most importantly, however, will be the global implications of the advent of AI. For the leading power of the twenty-first century will likely be the country with the most advanced AI technology. Furthermore, much of the social, political, cultural and economic power shifts of our era will be driven by the forthcoming technological revolution.
Big Brother Is Watching You
Stephen Hawking, the renowned British physicist, brilliantly summarized the main problems associated with AI: “Whereas the short-term impact of AI depends on who controls it, the long-term impact depends on whether it can be controlled at all.” Nevertheless, one of the primary challenges posed by AI lies much deeper. Instead of fearing rebellious AI that will someday supplant humans, we should be wary of AI that is always faithful to its creators.
AI algorithms, profoundly useful when in benign hands, can be very dangerous when used by malevolent actors. Imagine an authoritarian surveillance state that uses its omnipresent algorithms to control, monitor, and regulate the actions of its citizens. For instance, by using advanced AI-powered predictive analytics, such a system could predict the emergence of potential dissenters and quell any rebellion against future Big Brother in its infancy. In extreme cases, AI might mandate its citizens to wear biometric bracelets that monitor their internal biochemical processes. Combining this with analysis of social media activity and external actions, it might go under people’s skin and understand their thoughts and motivations. Furthermore, by regulating the flow of information in social networks, such types of AI might subtly change people’s perception of reality.
For instance, in the aftermath of the NBA-China scandal, Beijing allowed much nationalistic fervor to be expressed via social media against the NBA, until the government realized that this might damage China’s reputation abroad. Beijing then ordered its state-run media and censors to tone down their criticism of the NBA, thereby diverting people’s attention away from the issue. China’s “Great Firewall” is one of the integral parts of Beijing’s plan to completely control its population. China is now beginning to use even more advanced techniques than “mere” censorship. This is particularly true of the use of AI (social credit system) to control and repress Uighurs in Xinjiang. Arguably, worst of all, China’s model of digital dictatorship is not confined to the country’s borders and is being exported abroad.
Data is AI’s oil, and the more data you supply, the better AI algorithms are.
The advent of digital dictatorships will offer for the first time since the end of the Cold War non-democratic states an alternative to liberal democracy, sparking a renewed ideological competition. Of course, the rise of alternative ways of development could strengthen democracy; for instance, the need to deter an ideological competitor will encourage open societies to get their own houses in order and stand united against the new threat. Yet, the Second Cold War will be a new power struggle between the free world, led by the United States, and digital dictatorships, primarily China. Beijing is actively fostering indigenous technological innovation to become independent from American technologies and be able to build its own parallel Internet universe based on Chinese ethical standards.
Data is AI’s oil, and the more data you supply, the better AI algorithms are. In China, because of the lack of any privacy constraints, the government and Chinese tech companies are enabled to collect vast amounts of data from more than one billion citizens without any serious obstructions. Meanwhile in the United States, some politicians, such as Senators Josh Hawley and Elizabeth Warren, are considering leading efforts to break up leading American technology companies. This is despite the fact that such break ups will hinder these companies’ ability to compete with China on the global market, thereby diminishing Washington’s global standing and technological dominance.
Moreover, AI will likely exacerbate global inequality (the United States and China are the nations that possess the crucial combination of a vibrant startup culture, a large population, and leading technological companies). As such, it is likely that poorer countries—deprived of opportunities to evolve quickly via cheap labor and exports (as AI and robots replace humans)—will gravitate towards relationships with either the United States or China to stay afloat. Such dynamics will compound the formation of the bipolar international order and usher in another hazardous era of competition on all fronts, from culture and science to technology and ideology.
The Future of Liberal Democracy
Democracy is more successful than authoritarian dictatorships because of the former’s inherent efficiency when it comes to data processing. Instability in liberal democracies is distributed among many small units that individually do not threaten the existence of the system, while, in centralized states, volatility is concentrated.
Authoritarian states use centralized information processing, whereas democracies avail themselves of distributed processing. In today’s epoch, distributed processing has worked better; for instance, if one of the multiple processors (governments, central and local, businesses, citizens) failed, others were quick to return the system to equilibrium. In dictatorships, on the other hand, there is a single processor (authoritarian leader and his cronies) that might succeed, provided the ruling circle is comprised of intelligent professionals. But humans, even the smartest, are not infallible. Even a minor mistake can entail disastrous consequences.
As Yuval Harari has remarked, under the conditions of past centuries, dispersed data processing (democracy) proved to be more resilient. But the Roman Empire, Chinese dynasties, and many other states existed for centuries under centralized systems. Consequently, if conditions change again in the 3rd millennium, the balance of favor might tip towards centralized systems. After all, algorithms work better with more data, and in the centralized system, all data is concentrated in one processor. What used to be dictatorships’ main handicap (centralized control) might slide over into a decisive advantage in the new era.
Similar challenges await capitalism, a quintessential pillar of liberal democracy. Capitalism is more efficient than its competitors because it distributes the process of decision-making among many actors, which decreases the possibility of a single mistake collapsing the system. AI, however, might nullify this advantage of capitalism if centralized states begin to use AI algorithms in economic decision-making. For example, data collected about changing desires and preferences of citizens could be harnessed to mimic the “invisible hand” of the market to match supply and demand. Authorities could also use AI to detect potential vulnerabilities in the economy and cure them before a recession occurs. Business cycles could be eliminated. What is the point of free markets then?
In the interim, the AI industry’s natural tendency towards monopolization will eliminate another critical pillar of capitalism: free and unrestrained competition, while also deteriorating the living conditions of a newly formed “useless” class, as AI and robots eliminate jobs. The resulting drastic rise in the level of inequality will entail massive resentment among the “irrelevant,” who will become increasingly radicalized and susceptible to fringe ideologies and extremist views.
The Case for Open Societies
Undoubtedly, AI will undermine liberal democracy and empower authoritarian states. This is not, however, the reason to abandon the open society model and lament the fall of the free world. Because it is open societies’ ability to learn from mistakes, adapt to changing technological and societal conditions, and effectively address existing challenges through argumentative political discourse that make them resilient and flexible.
In other words, open societies are antifragile. “There is no stability without volatility,” wrote Nassim Nicholas Taleb in his 2014 book Antifragile. At first glance open societies appear less stable than their authoritarian counterparts, with their constant struggles for power and accompanying swings from one course to another as those in charge change. But in the long run, this volatility makes the system antifragile—or able to benefit from harm. Fragile things lose from external stressors, but antifragile objects improve their qualities as they experience disorder, stressors, randomness, or volatility.
Avoidance of small challenges makes the overcoming of larger difficulties more difficult—because small events teach important lessons on how to deal with similar problems in the future. Open societies learn from mistakes that make the system a bit unstable in the short term but much more resilient in the long run: just like deep learning algorithms.
Most authoritarian states, on the other hand, collapse when Black Swans, or unexpected and dramatic events, arrive, because dictatorships suppress volatility (e.g. dissenters, sharing of opinions, demonstrations and protests against the government) that would have otherwise helped them address these Black Swans by providing important information. As such, the scenarios outlined above that picture a gloomy future for democracy represent one of the many ways in which the course of events could unfold. It is impossible to predict the future of social events because they involve thinking participants. However, if we desire to address successfully the challenges posed by the advent of artificial intelligence to global stability and prosperity—if we want the world to be free and democratic rather than repressive and Orwellian, and if we want liberal democracy to perpetuate and avoid bloody social upheavals—we must strengthen the tenets of an open society and adhere to its underlying principles.
Sukhayl Niyazov is a freelance writer in Greece.