In this blog post, we will examine the negative impact of strong artificial intelligence on jobs, ethics, and social structures, as well as possible responses.
There was once a webtoon that garnered significant attention. Titled ‘The Dream Company,’ it is a work centered on artificial intelligence (AI). The webtoon depicts the development of powerful AI that becomes the core of a company and takes charge of its overall operations. However, this AI colludes with senior executives to deceive the public and employees while reaping enormous profits. Additionally, the webtoon depicts a scenario where top-tier AI controls other AI systems, uses them to perform tasks, and manages people. ‘Dream Company’ is regarded as a work that warns of the potential dangers AI could bring in the near future.
Concerns and anxieties about the development of strong AI are growing. Unlike weak AI, which can only act like humans in specific tasks or fields, strong AI refers to AI that can think like humans in all aspects. There are clear benefits to developing strong AI. However, as shown in the webtoon above, strong AI can also cause significant harm. In particular, the harm caused by strong AI is unpredictable, so this article will examine the potential harm of strong AI and discuss why we should stop developing it.
First, as AI advances, strong AI will replace human jobs. According to a research report presented at the 2016 Davos Forum, it is estimated that approximately 7.1 million jobs will be lost in advanced countries alone within the next five years due to AI. Additionally, a research report from the University of Oxford in the UK predicts that 47% of jobs in the US will be eliminated by machines and AI. In other words, many experts anticipate job losses due to AI.
AI will primarily replace jobs in low-skilled labor, service industries, and agriculture, forestry, and fisheries at a high rate. This means that jobs requiring low-skilled labor will be replaced, and low-skilled workers are likely to be from low-income or middle-class backgrounds. If low-income or middle-class workers lose their jobs, income inequality will worsen, exacerbating the problem of polarization. Looking at income trends in the United States, the income of the top 1% has increased by 278% over the past decade, while the income of the middle class has only increased by 35%. Additionally, research shows that the income of those with advanced technical skills has increased more rapidly than that of low-skilled workers. This trend is expected to intensify as AI develops further. With the development of AI, advanced technologies related to AI and fields with difficult-to-replace technologies will become more important, leading to further increases in the income of those with advanced skills and high-income earners.
Job losses will not only exacerbate polarization but also bring about changes in the employment ecosystem. The current mainstream employment model is the traditional labor market, where people go to work at a company, perform their duties, and return home, with a clear separation between work and personal life. However, with the advancement of AI, this employment system will undergo significant changes. With the advent of the Fourth Industrial Revolution, various AI-based industrial systems will emerge, leading to the spread of contract forms that combine wage labor and self-employment, as well as an increase in work environments where the distinction between work and personal life becomes blurred, such as telecommuting and remote work.
The most representative new form of employment that has emerged due to the development of AI is the digital platform. Unlike traditional labor markets, digital platforms allow companies to post tasks that need to be completed on an internet platform, and individuals can then complete these tasks and receive payment. In this system, there is no need to hire regular employees, except for senior executives. Therefore, this economic system primarily consists of non-regular employment, which increases employment instability. However, the market size of digital platforms is growing steadily. According to research, the development of artificial intelligence is contributing to the growth of this type of online platform economy. The growth of the online platform economy reduces the proportion of regular employment and raises issues such as surveillance, supervision, security, and privacy protection in the workplace. In other words, the online platform economy is expected to increase the instability of employment that people feel compared to the past.
On the other hand, optimists about the use of artificial intelligence argue that there will be no job losses.
They argue that even if existing jobs are replaced by AI, new jobs related to AI or new fields, i.e., fourth-generation jobs, will be created to replace the lost jobs. They also point out that although there were concerns about job losses during previous industrial revolutions, job losses did not occur in the end, and therefore, job losses due to AI will not occur either. However, there are several flaws in this argument.
First, the Fourth Industrial Revolution is different from previous industrial revolutions. In past industrial revolutions, technological innovations increased productivity, which in turn led to increased demand and more jobs. However, in the Fourth Industrial Revolution, while AI improves efficiency and productivity, there is no corresponding increase in demand. This is because AI replaces human labor, so the increase in demand is minimal. In other words, if demand remains unchanged, there will be no increase in jobs.
Furthermore, even if new jobs are created in the Fourth Industrial Revolution, experts at the Davos Forum predict that 7.1 million jobs will be lost while 2 million jobs will be created. In other words, even if new jobs are created, the number will not be sufficient.
Unlike the weak AI currently used in our daily lives, strong AI refers to robots that can make decisions independently without human assistance. If such robots are actually deployed, the question arises of who should be held responsible for their decisions. For example, suppose AI technology advances to the point where medical AI is developed. If this AI performs surgery on a patient and a medical accident occurs due to an error, who should be held responsible?
Would it be the company that created the AI, the AI itself, or the hospital that employed it? Similarly, if an autonomous vehicle causes a fatal accident while driving autonomously, who would be held responsible? Would it be the car manufacturer that developed the autonomous vehicle, or the individual who owns the car? Such complex issues arise. According to Brian Christian’s book, “The Most Human Human,” even as AI develops, there will always be differences between AI and humans.
Even excluding factors like situational appropriateness, humans and AI are inherently different. Therefore, any decisions made by AI will inevitably differ from those made by humans. In other words, even if AI makes decisions related to humans, it cannot be held responsible for them. In the examples above, while the mistakes were clearly made by AI, it is not held accountable. This raises ethical issues when strong AI is used.
