In this blog post, we will look at various perspectives on whether an AI judge can replace human judgment and justice.
The driving force behind a rapidly changing society: Automation
It is no exaggeration to say that humanity’s passion for automation has been the driving force behind modern science and technology. The desire to automate something has been a long-standing desire of humanity, to the extent that even the simple act of opening a door has been automated through a machine called an automatic door. Automation is a major change that causes confusion in the process of humans becoming gods and at the same time becoming slaves of automated systems, which is referred to as the Fourth Industrial Revolution and the hyper-connected society.
If the steam engine was the key keyword for the First Industrial Revolution, electricity was the key keyword for the Second Industrial Revolution, and the computer was the key keyword for the Third Industrial Revolution, moving towards an automated society, the key keyword for the Fourth Industrial Revolution is sensors and artificial intelligence. Sensors are responsible for peripheral nerves, and artificial intelligence is responsible for central nerves, forming a huge organism that crosses between humans and things. Recently, artificial intelligence has been proven to be possible with examples such as AlphaGo and Watson, and the range of applications is expanding as the performance of advanced sensors such as speech recognition is greatly improved.
Possibility of realizing the artificial intelligence of judges
The development process of artificial intelligence is quite simple compared to its performance. Implementing artificial intelligence is like teaching a machine a database. In other words, if there is a problem and the data for it, artificial intelligence can be implemented, and the amount and quality of data determine the performance of artificial intelligence. In the medical field, where there is an abundance of databases and a clear problem to be solved, artificial intelligence like Watson can perform so well that it can replace a human doctor.
The emergence of AI judges stems from this context. Numerous precedents and vast volumes of laws serve as a great database, and trials are a field where problem solving is clear-cut. In fact, in October 2016, the joint research results of University College London, the University of Sheffield, and Pennsylvania State University in the United States showed that AI judges could predict trial outcomes with an accuracy of 79%. Dr. Nikolaos Aletras, who led the research, said that the AI judge will help human judges by identifying patterns of judgment in complex cases rather than replacing them. However, it is clear that the ultimate goal of the research is to increase the accuracy to the point where it can replace human judges.
Roman Yampolskiy, a professor at the University of Louisville in the United States, argues that the question of “whether artificial intelligence can replace judges” and “whether it is desirable to replace the role of judges” are separate. As he says, it seems inevitable that artificial intelligence judges will replace fickle human judges.
Purpose of artificial intelligence for law, trials, and judges
According to Domitius Ulpianus, a Roman jurist of ancient times, law originates from justice and is the art of justice and equity. The weak are protected by law from the strong, and the strong are controlled by law so that they cannot bully the weak. Those who do not fulfill their obligations are punished by law, and those who do are guaranteed their rights by law. However, the law is implemented through trials, and the trial determines the sentence of the accused according to the law.
From this perspective, a judge must be rational enough to put aside all personal values and only look at the law, but since a judge is also a human being, he cannot completely remove his own subjectivity. Sometimes a corrupt judge makes a wrong decision for his own personal gain. However, since AI has no subjectivity, it can conduct trials “perfectly.” This is the motivation and purpose of replacing judges with AI.
Another reason for replacing judges with AI is to resolve the delay in trials. Currently, it takes months for a human judge to reach a verdict, but an AI judge can make a decision in a much shorter time. In addition, it can comprehensively use scientific and statistical tools such as lie detectors to determine the credibility of statements.
Is an AI judge just?
We have previously discussed how an AI judge could be implemented and why it is necessary. Now, we will discuss again whether the automation of trials with AI is desirable from the perspectives of quality, rights, and humanism.
Before discussing whether an AI judge has the qualities of a judge, let’s take a look at what qualities a judge has. As mentioned earlier, the law is a social norm set by the state, and the process of determining how much it has been violated and imposing punishment accordingly is a trial. So, would it make a good judge if one could memorize all the laws and reasonably determine the punishment based on precedents? A person who meets these conditions is not called a good judge.
A good judge is considered to be just. So, what is justice? This issue is so complex that it is difficult to cover in a single book, but let’s take a quick look at what a just judge is.
A just judge relies on the law but respects the common people’s concept of law. He or she empathizes with the case and feels compassion, and makes a judgment based on an understanding of social consensus. Therefore, a just judge listens to the verdict of the jury, understands the law well, and has an eye that looks beyond the law. This is because the law is not meaningful in itself, but a means to realize social justice. If it is found that the law does not reflect justice and fairness, the law should be amended immediately.
However, artificial intelligence judges lack this flexibility. To put it another way, artificial intelligence judges are students who are only good at memorization. Artificial intelligence can only learn results-oriented based on the actions of senior judges. Therefore, artificial intelligence judges who rely only on past precedents will not be able to respond to new types of cases or legal systems.
Obedience to an artificial intelligence judge
If a Brazilian were to pass judgment in a Japanese court, the Japanese would be averse to the ruling. This is because the judge would not understand the culture and characteristics of the region. The reason why humans trust judges is because they seem to empathize.
As Yuval Noah Harari mentions in his book Homo Deus, humans evaluate the guilt of others through empathy. An AI judge lacking empathy will lose this right. Even if powerful AI becomes mentally equal to humans in the future, humans will not completely trust AI. We have no basis for trusting and following their judgments because we cannot understand their thought processes.
Trials should be conducted by humans
In humanism, sovereignty has been treated as a very important issue. It is a universal truth that humans are their own masters and that this right is not transferable. Trials are an important process for dealing with conflicts in human society, and it is not desirable for non-human entities such as artificial intelligence to intervene in this process.
The automated society that humanity has dreamed of is a society that only does what it is told to do, but not a society that does the thinking for you.
