The dark forces of global politics—illiberalism, populist nationalism and protectionism—have recently reasserted themselves and threaten to subvert liberal democracy and the rules-based world order. These fundamental sociopolitical currents pose challenges to the democratic system. What is less often acknowledged or analysed, however, is the effect of technological change on politics.
Policymakers and politicians tend to disregard the importance of technology in defining the future of the sociopolitical system. The digital revolution has brought unprecedented challenges, which disrupt some of the foundational ideas of democracy. Exploring both the dangers and opportunities represented by AI is therefore essential to ensuring the future vitality of democracy and safeguarding the contemporary world order.
In the past, the effects of new technologies unfolded relatively slowly, allowing governments time to adjust. However, the pace of the ongoing technological revolution in artificial intelligence (AI) and robotics will not only be much faster but will accelerate over time. Unfortunately, regulators often do not fully realize the consequences of new technologies and adopt the necessary reforms until after the negative repercussions have become evident. The fourth Industrial Revolution will not be a single watershed event—it will be a cascade of tremendous changes, with disruptions occurring at an ever-increasing pace, forcing humankind to constantly reinvent itself in response.
Artificial intelligence involves the simulation of human intelligence by machines. Unlike conventional computer programs, AI algorithms (sets of instructions created to solve specific problems) learn by themselves—in other words, programmers insert data, which is then analyzed for trends and important inferences. Due to their sophistication and possible applications in almost all fields of life, these technologies are poised to dramatically alter the world order—and humankind’s primary aim must be to maximize their benefits and mitigate their negatives. A deeper examination of the implications of AI for liberal democracy is the key to resolving the issues we face today and will likely face tomorrow.
AI-Powered Information Bubbles
On 4 December 2009, an event occurred that went largely unnoticed, but which marked a turning point for the business model used by technology companies and will have far-reaching impacts on society. Google silently announced that it would use fifty-seven signals to customize search results. In other words, since 2009, there has been no standard Google.
AI-powered algorithms are widely used by leading tech companies—including, in addition to Google, Facebook, Amazon and Microsoft. These companies constantly collect data about our activity in order to tailor information to each person according to her interests, desires and preferences.
Initially, such algorithms were devised to help us avoid getting lost amid the flood of information. In an attention-scarce world, the personalization of the internet was the best way to ensure that information conforms to our unique personalities, allowing for an effective and convenient web experience.
However, personalization of algorithms has had negative side effects. Since personalization algorithms provide people with the information with which they are likely to engage, people soon segregate themselves into information bubbles, where their own beliefs are reinforced and they are not exposed to opposite views. This leads to confirmation bias, a distorted outlook on the world and dampened creativity.
“Left to their own devices, personalization filters serve up a kind of invisible autopropaganda, indoctrinating us with our own ideas, amplifying our desire for things that are familiar and leaving us oblivious to the dangers lurking in the dark territory of the unknown,” writes Eli Pariser, the internet activist who coined the term filter bubble. Information bubbles create virtual worlds consisting of the familiar: they are one-way mirrors reflecting our own views. In the words of Bill Gates, an information bubble “lets you go off with like-minded people, so you’re not mixing and sharing and understanding other points of view.”
Filter bubbles have far-reaching impacts on society and on democratic politics. Due to the echo chamber effect, certain complex or uncomfortable issues whose understanding is necessary for the proper functioning of democracy are removed from public discussion, since people are unlikely to pay attention to things that are either difficult to understand or unpleasant.
Filter bubbles also act as invisible propaganda (60% of Facebook users are unaware of the algorithmic curation of stories, to cite an example). They force people to believe that the information they consume represents the undisputed truth—and thereby amplify partisanship.
In the past, the truth was independent of the observer. But, in the age of information bubbles, the concept of truth is losing its universal character and being relativized and personalized—making it harder to reach consensus, since each side of a dispute now believes in its own infallibility.
When people are restricted to convenient ideological frameworks, they are less likely to reach political compromises and more likely to find themselves at odds with one another. Information bubbles fuel political polarization by facilitating the formation of identity groups, resulting in an inability to undertake collective action and ultimately fragmentation and even the total breakdown of society.
However, leading tech companies have already made substantial strides in tackling this issue. For example, in 2017, Facebook depersonalized its trending topics list, thereby widening its users’ exposure not only to themes that they are interested in and views that they support but to different perspectives and important global issues.
Big Data, Big Tech and the Tyranny of the Minority
AI not only affects individuals, however. The use of AI algorithms in politics both presents new opportunities and poses challenges to the democratic process as a whole.
Previously, politicians announced their views and citizens elected those whom they thought would best represent the interests of their country. Nowadays, the opposite is becoming true: politicians study the electorate’s views and adjust their own accordingly.
Candidates are increasingly using artificial intelligence to analyze data about voters’ preferences by researching their social media activity. When a candidate discovers that his platform is not a popular one, he can change it to conform to people’s demands. Democracy could ultimately be strengthened by the implementation of AI: analyzing databases would give a potential candidate a more precise image of what citizens need, thereby bridging the chasm between politicians and ordinary citizens and ameliorating the democratic process. As Pedro Domingos has put it, thanks to AI, “democracy works better because the bandwidth of communication between voters and politicians increases enormously.” There is a flip side, however. If a candidate’s convictions reflect AI research into voter preferences, democratic elections could turn into demagogic mass appeals, rather than the reasoned deliberation process envisioned, for example, by the Founding Fathers of the United States.
What is more important, however, is that the convergence of AI and political marketing will provide politicians with tools that not only enable them to examine voters’ wishes but to manipulate them. Using cleverly designed algorithms, politicians can study individual opinions and nudge undecided voters towards supporting them, through personalized advertisements. This, in turn, may lead to the tyranny of the minority. After all, if a candidate knows that certain citizens are her steadfast supporters, she will not focus her efforts on them—since she can instead target only swing voters, who will ensure her victory. This means that the fate of a country might be determined by several hundreds of thousands of swing voters, who will be barraged by propaganda from all sides and therefore unlikely to make rational decisions.
Liberty and Equality in the AI Era
Two of the essential pillars of democracy are liberty and equality. AI erodes both these principles.
As Y. N. Harari has noted, referendums and elections are always about human feelings, not rationality, because if democracy were a matter of rational decision-making, it would make no sense to give everyone equal voting rights, since some people are clearly more rational and knowledgeable than others, especially with regard to economics and politics. But, on election day, the vote of a winner of the Nobel prize in economics bears the same weight as the vote of an ordinary cashier. As Harari writes in Twenty-One Lessons for the Twenty-First Century,
for better or worse, elections and referendums are not about what we think. They are about what we feel. And when it comes to feelings, Einstein and Dawkins are no better than anyone else. Democracy assumes that human feelings reflect a mysterious and profound “free will,” that this “free will” is the ultimate source of authority, and that, while some people are more intelligent than others, all humans are equally free. Like Einstein and Dawkins, an illiterate maid also has free will, hence on election day her feelings—represented by her vote—count just as much as anybody else’s … This reliance on the heart might prove the Achilles heel of liberal democracy. For once somebody—whether in Beijing or San Francisco—gains the technological ability to hack and manipulate the human heart, democratic politics will mutate into an emotional puppet show.
AI will allow us to produce algorithms that could manipulate our thoughts and actions and, in the long run, understand us better than we do ourselves. Even today, we are increasingly putting our faith in algorithms—most people trust Google as their only source of knowledge, for instance. We will probably be forced to rely on algorithms even more in the future, because, amid the proliferation of political propaganda, algorithms will remain oblivious to manipulation, unlike ordinary voters.
Democracy has another defining characteristic: equality—social equality and equality of opportunity. Yet the implementation of AI might create the most unequal societies in history because data is gradually turning into our most valuable asset. And those who control data will control the future.
Since time immemorial, land has been the most important asset. The most powerful states of their times—such as the Persian, Macedonian and Roman empires—were also the biggest. However, since the Industrial Revolution, land has been diminishing in importance, while factories and machines have become a true source of power.
In the modern era, data will eclipse both land and machines as the most significant asset, and wars will be waged for control of its flow. Data will be the primary driver of the economy. Digital flows already have a larger impact on GDP growth than does the centuries-old trade in goods and services.
But, as with almost all technologically advanced sectors, a data-driven economy naturally gravitates towards monopolization due to algorithms’ reliance on data. This reliance creates a self-reinforcing cycle. If a company gains an early upper hand over its competitors, it will be very hard to resist the process of market concentration. With better services, it will attract more customers. More users will supply more data, while more data will improve AI algorithms. In turn, more advanced AI algorithms will attract even more customers, who will provide the company with additional data, and so on, ad infinitum. This explains why the tech industry is so heavily monopolized and why companies like Google, Amazon and Microsoft are totally dominant within their fields.
If left unregulated, the economy of the future will consist of monopolies and this will lead to an unprecedented concentration of power, as algorithms and robots based on AI eliminate most jobs, leaving millions of people unemployed. How will democracies cope with the demands of a useless class suffering from something worse than exploitation—irrelevance?
Communism and socialism emerged in the wake of the Industrial Revolution because the ruling class was unable to meet the demands of the working class. The new technological revolution is likely to usher in new ideologies and movements that will put pressure on democracy and try to exploit the resentment of the useless for their own benefit.
AI is therefore arguably the greatest threat to democracy, for it undermines its foundational pillars.
Will democracy survive these unprecedented challenges? I believe so. Democracy has a unique capacity to effectively tackle the issues of modernity, and its inherent flexibility and pragmatic resilience will allow it to adapt to the new realities.
Open societies, as liberal democracies are sometimes called, have an inherent advantage: they are reflexive and therefore able to effectively respond to challenges. The theory of reflexivity, simply put, states that our perception of events affects those events. Hence, since open societies provide avenues for unconventional critical thinking, they are better equipped to grapple with the challenges of modernity than their ideological competitors. As a result, even in the midst of unprecedented social changes and threats to their very existence, liberal democracies have managed to adapt and thrive, thanks to their openness to criticism and acceptance of reforms.
The dawn of an AI-powered technological era marks a turning point in the history of humanity. AI presents previously unimaginable opportunities that will help us reinforce liberal democracy.
Our ultimate objective must be to ensure the viability of democracy by encouraging a more active, non-partisan discussion of the implications of AI, in order to devise a blueprint for managing its effects. This will require enhanced multilateral cooperation, to avoid the zero-sum thinking that might lead to an AI arms race and exacerbated global inequality; a better public understanding of AI and politics in general, to ameliorate the democratic process; and, most importantly, wider coordination of efforts between the three pillars of the AI-powered world: governments, technological corporations and civil society. The future of democracy is in our hands.