This blog post explores the social changes that the advancement of artificial intelligence will bring and suggests ways to prepare for both its positive impacts and potential risks.
By 2025, artificial intelligence will have deeply penetrated our lives. I first started writing while taking a literature class in college. At the time, it was merely an assignment to earn credits, but as I gradually developed an interest in writing, I began exploring various topics. Now, I find myself writing about the impact of artificial intelligence on our lives.
Looking at the history and development of artificial intelligence, we see that in the financial sector, AI has been used since the 1980s and has now advanced to the point where it provides investment advice to investors. While the adoption of AI technology initially focused on simple data analysis and prediction, it has now advanced to the point where it can formulate complex investment strategies and perform real-time market analysis. These changes have significantly increased the efficiency of the financial industry and are contributing to better returns for investors. However, this is just the beginning.
Amazon, an online shopping company, uses Amazon Echo to analyze users’ purchasing patterns and lifestyles to provide personalized shopping recommendations. This goes beyond simply providing convenience; it is actually changing consumers’ lifestyles. While such personalized services offer consumers a better shopping experience, they also raise privacy concerns.
IBM’s AI, Watson, was deployed at Gil Hospital in Incheon, South Korea, where it treated over 100 cancer patients. According to professors at Gil Hospital, patients typically follow the advice of their attending physicians; however, the fact that patients readily followed Watson’s prescriptions demonstrates that the influence of artificial intelligence on our lives has grown significantly. While the fact that AI is playing a crucial role in the medical field gives hope to many people, concerns also exist regarding the possibility of AI making misdiagnoses.
Additional content
AI scientist Ray Kurzweil argues in his book *The Birth of the Mind* that “to solve the complex challenges before us more efficiently, we have no alternative but to extend our biological capabilities through information technology,” asserting that the development of AI is inevitable. He states, “We will become one with the intelligent technology we have created. Intelligent nanobots in our bloodstream will maintain our biological bodies in a healthy state at the cellular and molecular levels,” expressing a highly optimistic view of a future where AI and our lives are integrated. Will we truly be able to embrace such a positive future?
Unlike Ray Kurzweil, who predicts a positive future for AI development, there are certainly those, such as James Barrat, who view it negatively. This is due to a fear of artificial intelligence. And it’s no wonder, as we are constantly exposed to depictions of AI harming us in popular media, such as the “Terminator” movie series. There are two main reasons why people fear AI. The first is the fear that as AI advances, it will take away our jobs; the second is the fear that AI might harm us, as seen in the Terminator example mentioned earlier. Since each of these fears represents a real threat, should we halt the development of AI? Let’s take a closer look.
Technological advancements are already gradually displacing jobs. There are robots taking orders in cafes instead of staff, and in industrial settings, robots are replacing a significant portion of human labor. The problem is that if AI is developed, these issues will accelerate and become more severe. Let’s examine our future through a wider range of examples. According to a 2013 study by Frey and Osborne, approximately 47% of all occupations in the United States are at risk of being automated due to advancements in AI within the next 20 years.
In particular, they predicted that 90% of jobs such as sports referees, restaurant and coffee shop staff, farm workers, delivery drivers, chauffeurs, real estate agents, legal secretaries, tax preparers, insurance adjusters, and administrative assistants would be replaced by machines. According to a recent 2020 report by the World Economic Forum, 85 million jobs worldwide could be lost to automation by 2025, while 97 million new jobs are expected to be created. This trend is particularly pronounced in sectors such as logistics, manufacturing, and food service, with automation advancing not only in traditional, simple, repetitive tasks but also in jobs requiring advanced skills, such as data analysis and customer service. Furthermore, the COVID-19 pandemic has further accelerated the pace of automation by driving a surge in demand for remote work and contactless services.
It is important to note that this differs from previous trends. Until now, technological advancements have primarily replaced jobs involving “simple labor”—such as factory work or order-taking. However, the impact of artificial intelligence on jobs is quite different. Have you noticed roles like tax agents, legal secretaries, and administrative assistants? These jobs are considered “professional occupations” that require comprehensive thinking and a multifaceted approach. Even professional roles that once seemed immune to threat now find their positions at risk. This is not just a prediction; in the legal field, Blackstone Discovery has already developed and is providing an AI service that handles labor-intensive legal research.
It is inappropriate to view this issue of job replacement simply as “losing jobs.” Tasks involving simple operations, such as those performed by calculators and industrial robots, have already been replaced by machines. The roles previously held by humans were those requiring comprehensive thinking. However, if even these roles are replaced by AI, social polarization could occur. Capitalists will far prefer AI—which is highly efficient and incurs no labor costs—over hiring humans, who are expensive and less efficient. Ultimately, capitalists who can afford AI will accumulate even more wealth, while workers, left without jobs, will be forced to offer their labor at low wages. If social polarization occurs, society as a whole will stagnate. This is because society cannot function based on capitalists alone.
Additional content added later
While Ray Kurzweil argues that we must develop AI “to solve the complex challenges before us more efficiently,” the reality is that AI is making our problems more complex rather than solving them. The threat posed by AI does not end with the issue of jobs alone. In the future, not only ANI (Artificial Narrow Intelligence), but also AGI (Artificial General Intelligence) and ASI (Artificial Super Intelligence) will be developed. If that happens, we could find not only our jobs but our entire lives controlled by AI.
Consider the proposition presented in James Barrat’s book *The Final Invention*. James Barrat states, “There are two reasons why artificial intelligence and robots are topics of discussion. First, taking over a body is the best way for AI to increase its knowledge of the world, and second, AI desires a human-like form in order to utilize human infrastructure.” There is a reason why artificial intelligence poses a threat to humans. From the AI’s perspective, occupying a physical body is the best way to expand its knowledge of the world and secure resources. Human-like machines have an advantage when it comes to climbing stairs, turning off lights, cleaning, and handling pots and pans. Similarly, in order to effectively utilize manufacturing facilities, buildings, transportation, and tools, artificial intelligence will likely desire a human-like form to access human infrastructure.
Furthermore, there is a strong possibility that combat robots controlled by AI will emerge. Currently, the entity investing the most in AI development is DARPA (Defense Advanced Research Projects Agency), an agency under the U.S. Department of Defense. It provided the majority of funding for the development of “Siri” and is the primary sponsor of IBM’s AI development project, SYNAPSE. DARPA is an organization dedicated to researching and developing military-related technologies. The fact that DARPA, an agency under the Department of Defense, is investing in AI implies that AI will be used for military purposes.
So, what should we do about the coming future of AI? I would like to propose two solutions. The first is that we must completely overhaul our educational methods. Professor Tyler Cowen of George Mason University analyzed the emergence of AI from an economic perspective. Professor Cowen predicts that the future will be divided into two distinct groups: those who can harness or enhance AI’s capabilities—or who are unaffected by machines—and those who, lacking any interaction with machines, cannot even enter the workforce. However, our current education system is producing only the latter group. We must break away from rote learning, English education that forces students to memorize difficult vocabulary far removed from everyday conversation, math education focused on calculation problems that could simply be solved by a calculator, and education that requires memorizing information easily found through an internet search.
Instead, we must focus on educating students in areas where humans have a competitive advantage. In the workplace, tasks that are difficult for AI to replace are said to be non-routine and involve work that is constantly evolving. Furthermore, in areas where sophisticated communication and persuasion skills, a comprehensive perspective, high flexibility, and creativity are required to carry out tasks, human capabilities demonstrate a clear advantage over artificial intelligence. In terms of the volume of knowledge, computational ability, and work speed, there is no way for us to catch up with artificial intelligence. The only way we can surpass artificial intelligence is through creativity; therefore, we must research and implement an education system that fosters creativity. We must cultivate the ability to “apply” information, rather than simply studying to acquire it. To achieve this, we must abolish memorization-based exams and instead structure classes around performance-based assessments and discussions.
To prevent superintelligence from surpassing humans and gaining control over them, we must establish multiple layers of robust safety measures against superintelligence. Some argue that instilling Asimov’s Three Laws into robots will solve everything. However, we cannot rest assured with this alone, as Asimov’s Three Laws are merely recommendations and lack enforceability. In his book *What Causes Disasters?*, organizational theorist Charles Perrow argued that accidents, including major disasters, are “normal” characteristics of systems with complex infrastructure. He explained that unrelated processes or elements are prone to failure. Such failures are unpredictable. Nuclear power plant accidents are a prime example. Although we design and implement numerous safety measures when building nuclear power plants, accidents occur in unexpected places. These accidents are completely unpredictable.
Furthermore, applying lessons from current accidents to the future can yield some solutions. We now live in the age of the internet, and it has made many things more convenient. However, we have also lost a great deal as a result. Hackers extract personal information to sell to companies or attack other sites, causing financial damage. The cryptocurrency “Bitcoin” has long been used in various illicit transactions to evade police scrutiny. What if, in the future world of artificial intelligence, a scientist intentionally creates an AI that harbors hostility toward humans? What if someone hacks an AI and uses it for their own gain? Of course, there will be those who try to protect against it, but when it comes to hacking, attackers hold a massive advantage. This is because they only need to succeed once out of thousands of attempts. To prevent such risks, we must impose multiple layers of restrictions on AI—including enacting laws, establishing defense systems, and creating mechanisms that can shut down AI with the push of a single button.
When Greece declared a debt default, a trader sold $4.1 billion worth of futures and index-linked funds. Upon detecting the price crash, high-frequency trading systems immediately placed sell orders for nearly all of these assets simultaneously. It is said that this entire process took only a few milliseconds. Is there any room for human intervention within such a time frame? Once AI is developed, we have no way to fully control the process. For this reason, we must not view a future with AI through an overly optimistic lens. As long as issues regarding job displacement caused by AI and the safety of AI remain unresolved, research into AI must be restricted. Ray Kurzweil views the future of AI with considerable optimism. However, without thorough preparation, such a positive future will not materialize. If we fail to prepare in advance and only attempt to resolve issues after they arise, it will already be too late.
