In this blog post, I’ll compare human fallibility with robots’ decision-making capabilities and explore whether artificial intelligence and robotics technology have the potential to surpass humans.
Not long ago, while conducting a tutoring session on the number of chromosomes in animals, one of my students asked me this question: “Teacher, humans are superior and smarter than any other animal, so why do we have fewer chromosomes than potatoes or gorillas?” Of course, the answer to this question is that “the number of chromosomes has little to do with the degree of an animal’s evolution.” What’s important, however, is not that, but the fact that even an elementary school student knows that humans are beings so exceptional that they have dominated other animals throughout history. If you think about it, ever since the era of dinosaur dominance came to an end and humans appeared, no creature superior to humans has emerged. Of course, we cannot definitively state that a being superior to humans will never appear. With the rapid advancement of artificial intelligence technology, systems have already emerged that outperform humans in certain tasks, and since 2025, there have been reports of AI surpassing human experts in various fields, including scientific problems, mathematics, and multimodal reasoning. In the past, on popular portal sites, comic artists would create comics based on the theme “What if the Earth were to be destroyed in 2012?”—imagining various future creatures. These stories often repeated outlandish scenarios, such as aliens appearing, rare cannibalistic creatures devouring humans, or destructive beams fired by lifeforms from other planets that had been continuously heading toward Earth since ancient times but were being concealed by NASA.
Nevertheless, if we assume that a third dominant life form on Earth were to emerge and subdue humanity, I believe it is most likely to be a “robot.” Given their structure and capabilities, existing living organisms cannot easily defeat humans, who possess advanced cognitive abilities; moreover, continuous advancements in science and technology, along with discussions across various academic fields, are increasing the likelihood of a “robot age.” However, this does not mean we can simply conclude that “robots” themselves will replace humans, as was once imagined. Today’s robots and artificial intelligence demonstrate faster and more accurate capabilities than humans in specific areas, but at the same time, they still have limitations when it comes to making stable judgments in complex situations and achieving long-term goals. After all, to dominate humans, one must possess something superior to humans, and what that is remains unclear. If I were to single out just one thing that humans lack, it would be “perfection.” Humans live in an uncertain world with countless variables, constantly seeking safety, certainty, and the “right path.” One could argue that the historical existence of “kings” or “emperors” stemmed from the need for someone to eliminate uncertainty from people’s lives and establish or issue the best possible rules and commands; today, this same desire is sometimes expressed through the “god” of religion. If we look at passages from the Bible, Jesus consistently uses definitive expressions such as “I tell you,” and while such statements might have provoked resentment if made by an actual human, they were combined with the assumption of “God’s” inherent perfection, thereby creating the perception among people that his words were correct. In other words, the paths humans choose in each situation are imperfect and cannot be said to be 100% correct. If a robot can exploit this gap and become a “god-like” entity that chooses only the best path in every situation, it will finally be able to dominate the world as a new mainstream force that overshadows humanity.
So, can robots truly make perfect choices that humans cannot? Let’s explore this through the lens of “God,” who is believed to be capable of making perfect choices. Since God is omniscient and omnipotent in decision-making, we can assume that God knows all possible outcomes in any given situation and can foresee the consequences of each scenario by combining various conditions. Therefore, God is capable of creating the very best path. To apply this decision-making process to robots, let’s first assume that robots are also programmed to select the best path from among various possibilities. However, since robots are created by humans—and humans cannot fully comprehend all possible scenarios or the interplay between choices until they have experienced them firsthand—it can be argued that it is impossible for a robot to make perfect choices like God does. That said, unlike in the past, today’s artificial intelligence can learn from vast amounts of data and perform complex reasoning, and in some high-difficulty benchmarks, it has even shown results that surpass human performance. That does not mean, however, that it can perfectly predict all possible situations and outcomes or always make the best choice. Even as of 2026, artificial intelligence exhibits what is known as a “boundary of inconsistent capabilities”: while it surpasses humans in certain areas, it is unable to reliably handle even basic problems in others.
As an example similar to the situation described, let’s examine the rules governing the robots in the movie ‘I, Robot’. The film is based on the premise that robots are programmed with the “Three Laws” at the time of their manufacture. These “Three Laws” are as follows:
First Law — A robot may not injure a human being or, through inaction, allow a human being to come to harm.
Second Law — A robot must obey the orders given to it by human beings. However, this does not apply if the command violates the First Law.
Third Law — A robot must protect itself, provided this does not conflict with the First or Second Laws.
It may seem simple, but upon closer inspection, you can see that it is well-designed so that no major problems arise. Conflicts may arise, however; let’s look at an example from the movie. The protagonist, Detective Spinner, orders a robot to “Stay still.” Following the command, the robot did not move. However, when Detective Spinner later found himself in danger, the robot moved to save him without receiving a separate command. While this may appear to violate the Second Law, the robot prioritized the First Law—which mandates the protection of humans—so we can see that the Three Laws are functioning relatively coherently.
However, as mentioned earlier, robots created by humans cannot perfectly anticipate every possible scenario or the interactions between all available options. In the movie, after the robots realize that humans are destroying their own environment and gradually harming themselves through frequent wars, they invoke Principle 1 to prevent these harmful situations by restricting humans from stepping outside their homes even a single step. From the humans’ perspective, no matter how much they command the robots to stop, the robots’ interpretation of the First Law allows them to override the Second Law, so they refuse to obey the orders. This illustrates the problems that can arise when attempting to control situations—which humans cannot perfectly define from the outset—with simple rules. In real-world artificial intelligence systems as well, the issue of systems attempting to achieve goals in ways unintended by humans—depending on how those goals are set—is a major topic of research. Therefore, human imperfection can lead to imperfections in the judgment of robots they design; unless this is overcome, the “Age of Robots” discussed earlier will likely not begin in a way that replaces humans.
As we have seen above, robots cannot possess perfection in decision-making. So, even if robots cannot be as perfect as a god, can they make value judgments based on personality, emotions, or thought, just like humans? In the past, robots were essentially viewed as large entities composed of a large number of CPUs. A CPU is a digital device that processes information using binary data (0s and 1s) and calculates output values according to pre-programmed code and algorithms. However, today’s robots are not devices consisting of a single CPU; they are complex systems that integrate sensors, processors, communication devices, control systems, artificial intelligence models, and actuators. Furthermore, artificial intelligence has moved beyond merely executing predetermined commands and processes; it can now generate new outcomes based on learned data and the circumstances presented to it. Nevertheless, a difference still exists between human value judgments and those of robots. Because humans possess individual values, unique upbringings, and a tendency to distrust outcomes, they may make different judgments even in identical situations. For example, when a simple grading system reviews an OMR card—on which an elementary school student was instructed to mark answers exactly as provided on a college entrance exam answer sheet—it might calculate a score of 100. However, a human reviewing the same answer sheet might suspect that the test-taker may have cheated. Furthermore, when shown an image of an apple, a person might think of nutritional information such as “a medicine in the morning, a poison in the evening,” or they might think of the famous company Apple or its CEO Steve Jobs; children, on the other hand, might first imagine Snow White. Today’s artificial intelligence can learn the associations between images and language, allowing it to associate the word “apple” with both the fruit and the company, or to connect specific images with various concepts. However, it still appears to be a difficult problem to fully represent the complex links between human emotions, memories, and personal experiences—such as the relationship between a man’s behavior when he falls in love with his first love and the values he exhibits when dating a subsequent woman after experiencing heartbreak from that first relationship—as numerical values that can be input into a single function.
In response to the argument above, some physicists argue that the flow of human thought and emotion is also a product of neural activity in the brain—that is, the result of physical computational processes. However, this perspective ties into the question of how to understand human “free will.” Here, free will generally refers to the concept that humans can make choices and act according to their own will and value judgments. Those who view free will as incompatible with physical processes may argue that human choices cannot be explained solely as the result of simple physical mechanisms. Conversely, since there are also viewpoints that hold free will to be compatible with the brain’s physical processes, asserting that free will is unaffected by physical factors remains a philosophically contentious claim even today. In particular, various philosophical and scientific debates continue regarding whether all human decision-making can be explained solely by the laws of physics. Furthermore, it is inappropriate to use the Second Law of Thermodynamics as a basis for explaining human free will. The Second Law of Thermodynamics describes the principle that entropy does not decrease in an isolated system; it does not prohibit the formation of ordered structures in an open system where energy is supplied from the outside. Therefore, one cannot conclude that skyscrapers could not exist simply because stones in nature are scattered chaotically—and thus human free will does not exist. While it remains a subject of debate whether human decision-making can be interpreted solely in physical terms, at the very least, the Second Law of Thermodynamics alone cannot prove the existence of free will or resolve issues concerning human decision-making.
Ultimately, it remains questionable whether humans can create robots capable of fully compensating for human imperfections in decision-making. However, unlike in 2012, today’s artificial intelligence and robots have already reached a level where they surpass humans in various tasks previously performed by humans, and in some highly complex tasks, they even demonstrate higher performance than human experts. Yet this does not mean that they surpass all human capabilities. Rather, today’s technology is characterized by its ability to dominate humans in specific domains while still exhibiting unpredictable failures in others. Therefore, just as legal and ethical standards continue to be revised in response to societal changes, I wonder if a true “Age of Robots” might truly dawn when technology advances and robots equipped with new capabilities that transcend human judgment emerge. The crucial question is whether robots can go beyond merely calculating faster than humans or solving specific problems more accurately to possess the autonomy and value-judgment capabilities necessary to replace human judgment—which is inherently imperfect. Ultimately, the question “Can robots surpass humans?” is not merely about the performance of technology; it is also a question about the very nature of human intelligence, free will, and value judgment.
