Can a collective or single government body be replaced by artificial intelligence? Traditionally, in our country, presidential elections are the reason for political bargaining on another change of the political regime along with strengthening or weakening the president or the parliament.
At the same time, as it turns out, sometimes this pendulum can swing toward excesses such as authoritarianism. In this context, a political system is considered as a plasticine product that can be of any shape, depending on the current balance of main political players.
Such political bargaining is practically not analyzed by the academic community in detail, since neither history, nor philosophy and political studies do not provide a clear answer to the question of how to give structure to authorities. Lee Kuan Yew, whose smiling face can be found on the shelves of almost every Ukrainian bookstore, disturbs the judgment of Western liberalism supporters insistently. Meanwhile, ‘strong hand policy’ apologists require to be constantly reminded of the tragic history of the USSR formation and disappointing trends in the development of our neighbors.
However, putting aside historical parallels, solving the problem of appropriate concentration of power largely depends on the answer to the key question: who has the highest faculty (and can more likely) to make appropriate and effective decisions: a collective body or an individual?
Both ancient Greeks (for example, Aristotle believed that a group made decisions at least not worse than its best member) and the collective wisdom of the Ukrainian people (four eyes see more than two; so many men, so many opinions) also tried to answer this question.
However, the more or less reliable answer has been found only in the last few decades, when professional psychologists made the judgement. Just then it turned out that, as it often happened before, neither people nor Aristotle had any idea what they were talking about. But let me explain everything step by step.
Kraken in power
How likely is that a sole state leader will make appropriate decisions? First of all, we should immediately reject the option of a truly one-person decision-making process by such a leader. The fact is that when making a decision, one person is able to effectively analyze only a small number of factors. As the problem becomes more complicated, the analysis no longer makes sense, and here you should rely solely on your intuition. Given the complexity and diversity of areas that a potential authoritarian leader has to control, he is not able to govern the state at minimally acceptable level alone even theoretically (if, of course, this state is not an island inhabited by dozens of aborigines).
Therefore, we will leave alone assertive leaders-dictators, but take a look at another less exotic decision-making type – making decisions by hierarchical groups. The main point of such a group is that problems and possible ways of their solution are discussed by a rather wide circle of advisors, but their leader takes a decision alone after the group discussion. The aforementioned model reflects an ‘informed dictatorship regime’ more adequately in a modern complex world. So, does this model have any drawbacks?
As it turned out, it has a lot. Aside from potential individual weaknesses of a single-person leader (a lack of perception, personal development, education or experience), there are still a series of mistakes that almost all the leaders make.
To begin with: which of the advisors, in your opinion, will a leader listen to first of all? Perhaps, the most intelligent one?
Usually an advisor who demonstrates more confidence in his advice is the most likely to succeed. This rule works even when such confidence is in no way correlated with the reasonableness and value of the advice itself (that is, when confidence is groundless).
However, even an advisor who is bursting with confidence has a decent competitor. Who is this influential person? Maybe the most experienced advisor? Again no. A leader’s personal opinion has one of the most important priority in the leader’s coordinate system. As a rule, the influence of the leader’s personal opinion on making decisions is practically equivalent to the position of a professional expert. Even if such an expert has much extensive knowledge and experience in the relevant area.
There is the opposite effect when a leader doesn’t have an opinion regarding problems. In this case, leaders tend to take the advice of an advisor even when this advisor has less knowledge and experience in the relevant range of problems than the leader himself.
Another typical phenomenon is the negative natural selection among advisors of a sole leader. As researches found out, advisor’s advices, which turned out to be false, greatly limits its influence in the eyes of the leader. At the same time, the advice, which turned out to be correct, raises the advisor’s profile in the eyes of the leader just a little bit.
That is, it is too difficult to establish credibility with the leader, but one can lose it as quick as lightning. As a result, the leader accumulates advisors guided by the strategy of giving vague and dubious advices, which would not give rise to accusations in the future, around himself.
In addition, if the leader is really seeking advices, he tends to trust the advisor who previously agreed with the leader’s opinion more. And this means that over time, such a leader will inevitably fall into the warm bath of his illusions supported by the persons closest to him.
In addition, to be successful, both the leader and his team should have developed cognitive skills and consciousness at the same time. If the leader has no such qualities, it can not be compensated by any team of advisors. To the contrary, a genius with a narrow-minded team will take mostly wrong decisions.
Finally, getting the power significantly changes a person’s behavior. As Dacher Keltner’s studies described in detail, a person who receives power tends to behave more impulsively and selfishly, lie, violate the rules, which, in their opinion, other people should follow, and even… take away candies intended for children (I understand, it is funny, but that’s a fact).
It would seem that supporters of a strong parliament have something to celebrate. Their suspicions that all historical horrors of authoritarianism are not a temporary domestic specific features, but a scientifically verified pattern are confirmed.
However, are there any significant benefits of collective decision-making processes?
All power to councils?
In terms of the optimality of solutions, such advantages turned out to be not so obvious.
First, collective bodies usually not only do not remove the mistakes of their members, but, on the contrary, aggravate them.
Secondly, making decisions by a group entirely depends on absolutely random factors. For example, the position of that person who first expressed his opinion is often crucial.
Thirdly, if members of any collective leaned toward the same idea before the discussion, the result of the discussion would be only a significant increase in their confidence about such an idea. As a result, the collective is radicalized for no apparent causes.
Fourthly, groups assign high priority to information, which is known only to individual members of the collective. This disadvantage is especially obvious when the time to make a decision by the group is limited to a small period.
In addition, it is ironic that members of the group, who did not have the opportunity to discuss a problem, often make a conclusion, which is much closer to the truth than the conclusion of the group that discussed this problem.
Finally, the nearly main drawback of a collective decision-making process is its slowness. For example, a parliament rarely considers a draft law faster than in 6 months. This slowness is completely in contradiction with modern trends: in the private sector (even when it comes to huge corporations), every hour, minute and sometimes second counts to make decisions. A company that can not make decisions quickly is lost in the competitive market in the blink of an eye.
In terms of quickness, a sole leader can teach any parliament a thing or two. Although, let’s be clear here, sometimes failed states have the rulers who wait to the last minute to make decisions because of their personality features.
It is all over!
It is easy to see that both collectives of equal members (voters in the referendum, parliament), and hierarchical groups (sole leader and his advisers) themselves are imperfect and programmed to make mistakes. Therefore, in terms of accuracy and efficiency of decisions, the problem is not the one who makes decisions (hierarchical or equal group), but the way, in which the works of such groups is organized.
For example, the involvement of a person who actively advocates a position different from that of the majority has a positive impact on making adequate decisions by the group. At the same time, if the group understands that its member took such a position only due to his status as an official opponent, any positive effect of such participation is rapidly falling. Therefore, there is a lesson for the parliament: opposition factions have a positive influence on the quality of decisions made by the parliament only if the reason for their opposition is a real analysis of initiatives, but not a denial as a result of the political struggle without reference to the essence of the problem.
The general conclusion is that without adequate discussion procedures, targeted formation of groups and thoughtful analysis of the expertise level of their members, both hierarchical and equal groups will be able to make optimal decisions only by positive coincidence.
However, the problem is that groups (both with and without a leader) are able to organize themselves by their own rarely. Often the groups do not appreciate the opinion of the best expert in the group while overestimating the least-informed advisor.
I’m from Zuckerberg
If a person is pre-programmed to make false decisions, probably, it is time to replace them with something else.
It is projected that in the next decade, people may stop writing programs, since artificial intelligence is quite capable to cope with coding tasks. And here it is worth mentioning that both program codes and laws have a lot in common: the one and the other is the algorithms, rules of conduct. Moreover, a program code is even a little more complicated, because in the case of an error, any program definitely stops to work correctly. And in the case of a mistake in the legislation, there is always a chance to deviate from the literal interpretation of provisions and use many other types of interpretations (systemic, teleological, historical, etc.) that would give the meaning even to a completely inefficient law.
Indeed, in recent years, making decisions is delegated to artificial intelligence with increasing frequency. Some innovative companies are already including it into the corporate bodies, and government bodies are introducing it in the decision-making mechanism based on the processing of large amounts of data. As for the justice system, artificial intelligence can predict the behavior of perpetrators and choose appropriate penalties or evaluate the probability that suspects will not hide from the investigation. Chat bots consult and response to citizen inquiries and even deal with their recruitment for military service more successfully than people. In this case, we are talking about already working algorithms. There are even more areas, in which the algorithms are ready for use, but not yet allowed by people. They cover verifying the grounds for providing social assistance, making decisions on granting a preferential loan to a beginner businessman, issuing a visa or an immigration permit, selecting entities for inspections, etc.
The next logical question is: can a collective or single government body be replaced by artificial intelligence?
In the context of choosing the best solutions, artificial intelligence seems to be perhaps the best option. The algorithm is not burdened with obligations to party partners, does not think about re-election for a second term and is entirely deprived of human’s cognitive defects.
But humans would not be humans if they were looking for simple ways. Therefore, we have another problem – researches indicate that we, humans, do not like when algorithms make difficult decisions for us. Even if such decisions are more optimal and effective. Even if such a difficult decision turned to be false in terms of consequences, it often does not prevent people from feeling completely satisfied.
As for government bodies, it became clear from the above researches that they are not always a place for making optimal decisions. Typically, they are generators of decisions, which are a product of political (often situational) compromises. Although such decisions are not perfect, they, like any compromise, are more readily perceived as compulsory.
However, this conclusion does not mean that algorithms do not have a place in the government system. Artificial intelligence can safely take the lion’s share of responsibilities of government bodies related to secondary or preparatory works: data collection, answers to requests for information or assignment of tasks among subordinates.
But it hardly needs to be limited with these aspects. It is likely that in a few years, artificial intelligence will become a full parliament or government member in any country.
Such a parliament or government member will have a larger political significance that its ordinary human colleague has. The algorithm can act as a benchmark, deviation from the proposals of which requires detailed justification and argumentation. Moreover, discussions with the participation of artificial intelligence would be hardly limited to straightforward political fights while shifting government or parliamentary discussions from unstoppable promotions to discussing the essences of problematic issues.
In this scenario, advertising slogans of software developers such as “It works”, “Because it’s consistent”, “This hardware did not steal anything” or “Take and do” will be a clear indication of inevitable changes.
The opinions expressed in this article are just the author’s own