In this blog post, we’ll examine the limitations of democracy—namely, human irrationality—and consider how democracy might change if artificial intelligence were introduced into voting and political systems.
An upset occurred. In the recent U.S. presidential election, Republican Donald Trump secured a majority of the Electoral College votes, confirming his election as the 45th President of the United States. It was a result no one had anticipated. Major opinion polls had unanimously predicted a victory for Democrat Hillary Clinton, but the outcome was unexpectedly a victory for Donald Trump. Who is Trump? He has called for a ban on Muslims entering the United States, compared Mexicans to rapists, and mocked people with disabilities. And that’s not all. He has faced allegations of sexual harassment and hurled insults at women who claimed to be victims, suggesting they weren’t attractive enough to be sexually harassed. He also claimed that a judge with Mexican parents could not be impartial, and during a presidential debate, he disparaged his opponent Hillary Clinton’s appearance and said he would send her to prison. Beyond that, the controversies surrounding Trump were endless. Viewed rationally, Trump was a disgraceful candidate. He was a man who championed hatred and bigotry as values. He was a symbol of extremism and the worst possible candidate. Even voters who concealed their support for Trump in opinion polls knew he was a disgraceful candidate. Nevertheless, American society—a democracy—elected this worst of extremes as its representative through the democratic process of voting. Democracy chose him. For many, it was a night of despair. As seen not only in the results of the U.S. presidential election but also in various other events—including Brexit in the UK and the impeachment of President Park Geun-hye—democracy is far from perfect.
Of course, it is certain that democracy promotes fairness and participation in the governance of a nation. Democracy does not discriminate based on race or region; it gives the poor and the rich equal opportunities to voice their opinions. Even the poor, who are socially disadvantaged, can play an important role. Crucially, democracy preserves individual freedom, equality, and dignity. That is why democracy is regarded as the best form of government today, and the majority of nations are adopting it.
However, this framework of democracy is giving rise to irrational phenomena, such as the incidents mentioned earlier. In many respects, democracy is failing. Why exactly are these phenomena occurring? The reason is that we make irrational judgments based on our emotions and are biased toward what we want rather than what we need. Rather than making decisions based on statistics and scientific analysis, we often fail to exercise rational judgment, reacting instead to external emotional stimuli. As seen in the Asch Experiment conducted by the American psychologist Solomon Asch, humans tend to make decisions based on emotion rather than objectivity. The more complex the issue, the more likely we are to make irrational, emotion-driven decisions rather than relying on various disciplines such as statistics, sociology, and psychology. We cannot be experts in every field, nor do we have the time to analyze them all. From this perspective, there are limits to human judgment and knowledge.
In the 1950s, an American psychologist named Solomon Asch conducted a series of experiments to identify the human tendency to conform to the actions of others. He had a test subject enter a room where nine chairs were arranged in a semicircle and had them sit in the second-to-last seat. Eight people then entered one after another and sat in the remaining chairs. Unbeknownst to the test subject, they were Asch’s assistants. Asch showed two cards in succession. The first card had only one line drawn on it. He then asked each person, one by one, which of the three lines on the second card was the same length as the line on the first card. Since the three lines differed by as much as 2 inches, it was not a difficult task at all. However, by the time it was the test subject’s turn to answer—the eighth person—the other seven had already given their opinions. To the test subject’s bewilderment, all seven had chosen the wrong line, and what’s more, all seven had agreed on the same incorrect answer. A conflict began to arise between the subject’s visual judgment and the consensus of the other seven. Which should they believe? Twelve out of the eighteen subjects followed the majority opinion and chose the wrong answer. However, when asked after the experiment whether they had been influenced by others, they unanimously replied that they had not. They had not merely followed what others were doing; they had, naively, changed their own beliefs.
A democracy in which such irrational people participate tends to favor the “fairest” decision—the one that garners the most votes—rather than the best decision, regardless of whether that outcome is right or wrong. So, what if we introduced artificial intelligence into the democratic system to make the most rational and reasonable decisions? You’ve probably heard that artificial intelligence will eliminate many jobs. But could AI even take on the role of government? In fact, in Silicon Valley, many AI teams are working to replace members of Congress, the legislature, and ultimately, traditional politics. Using blockchain—a database technology designed to prevent tampering—they have already created virtual parliaments and virtual nations, and are developing more efficient voting systems.
If cutting-edge AI technology were introduced into the voting process, it could improve democracy in several ways. First, through big data and machine learning, AI could understand individual tastes and preferences and assist voters. It would help voters easily and thoroughly familiarize themselves with the facts before making a decision, enabling them to make more rational and reasonable choices. Crucially, it will accurately assess biased media coverage and partisan arguments, thereby reducing instances of emotionally driven decision-making. It is said that after the UK’s Brexit referendum, the second most frequently searched question on Google by Britons was “What is the European Union?” This suggests that many voted to leave the European Union without even knowing what it was. At the very least, if AI is properly implemented, decisions made out of such ignorance will cease to exist.
Furthermore, artificial intelligence—which can work tirelessly—will immediately detect system errors or malfunctions, propose solutions, and make the political system more robust. If trust in the government falls below a certain threshold, the system could be configured to recognize the situation immediately and trigger new elections. In Taiwan, artificial intelligence technologies such as machine learning are being utilized to facilitate communication between the government and citizens and ensure transparent information disclosure. If artificial intelligence is fully implemented, it will be able to represent the opinions of all citizens more transparently, accurately, and swiftly. The history of representative democracy, which has spanned several hundred years, will come to an end, and a new era of direct democracy will dawn. AI will be able to represent the will of the people and allocate laws and budgets for their benefit. Until now, politicians have engaged in back-and-forth negotiations, delaying the passage of bills for years; however, AI will be able to process bills that used to take years in just a single day.
Furthermore, AI could potentially replace the current “one person, one vote” system with one that awards bonus points to those with more information or a higher level of education. In other words, it will be possible to reflect the opinions of experts more fully to achieve more desirable outcomes. If such a system were introduced, there might be fears that the highly educated upper class would steer society in the direction they desire. However, this is not necessarily the case, as there are certainly experts who represent every social class. For example, someone like Bernie Sanders in the United States is highly educated yet advocates for people in the lower socioeconomic classes.
Under the current democratic system, people from all social strata are granted equal voting rights, but many from lower socioeconomic classes are unable to actually choose the legislation they need. It would be far more effective for experts—who understand what legislation these lower-class individuals truly need and can advocate on their behalf—to make the optimal choices for them.
American futurist Thomas Frey argues that as artificial intelligence technology continues to advance, fixed election dates will eventually disappear, and issues will be automatically addressed and voted on as needed. Ultimately, all elected officials will be eliminated and replaced by automated voting by the general public. Perhaps artificial intelligence will create an automated consensus system, rendering voting itself obsolete. Artificial intelligence will be able to learn what we need and how we think, and automatically vote on our behalf. Since AI developers today are adopting the strategy of “don’t model the world; model the human ‘mind,’” this is a scenario that is entirely plausible in the near future.Democratic systems are destined to change. With technological advancements, the likelihood of such changes has increased even further. While it may seem risky for AI to intervene in the realm of politics—which is rooted in individual values—our values must be grounded in the facts of social phenomena. AI will be the very bridge connecting those facts to our values. It will aid humans in making rational decisions and further refine the role of government. AI will rewrite parliamentary rules and reshape the very structure of government.Some readers may wonder whether these changes will ultimately destroy democracy. If voting is automated by AI and it begins to assign bonus points to certain opinions, this might conflict with the very definition of democracy. However, improving democracy and destroying it can be very similar processes. Considering the social and economic costs we have borne due to democracy, such changes are inevitable. Perhaps technological development could even overturn the very concept of democracy. Of course, in this process, we need to deeply consider where human dignity ends and where AI is permitted to intrude. One thing is certain: by making more rational and logical decisions through AI—which produces balanced, optimal outcomes rather than simply reflecting the majority opinion—we will be able to build a more perfect society.
