The whole world is talking about artificial intelligence (AI). Governments speak of national competitiveness. Companies speak of productivity. Schools speak of AI education. The political sphere has belatedly begun to speak of regulation and promotion together. Yet the most important question has not yet been adequately addressed.
"Will AI make humans freer? Or will it make them more exhausted?"
On the 25th, Pope Leo XIV raised the question of human dignity in the AI era. Just as labor became a civilizational agenda during the Industrial Revolution, his concern is that human time, judgment, labor, and dignity are becoming the new social agenda in the AI era. AI is not merely a technology. It is a force that changes society's operating system. Whether that force becomes a tool that helps humans, or a device that subordinates humans to ever-faster systems, has not yet been determined.
In "The Burnout Society," Byung-Chul Han viewed modern people not as subjects who obey commands, but as achievement-subjects who believe they must produce results on their own. He called this an "achievement society" because the outer shell of freedom ends in self-exploitation. AI could become the most powerful amplifier of this achievement society. In the past, the phrase "there isn't enough time" was the minimum line of defense. Now even that line of defense may collapse. "We have AI, so why isn't it done yet?" This phrase could become the new discipline of the workplace.
The claim that AI reduces work is only half true. In real organizations, AI-generated drafts trigger just as many demands for revision. Faster reporting is required. More granular performance measurement becomes possible. The volume of work one person must handle does not decrease. Only expectations rise. If that happens, AI is not a technology that lets humans rest. It becomes a technology that automates the burnout society.
That is why discussion of AI cannot remain within industrial policy. Semiconductors, data centers, and talent development are important. But that alone is not enough. AI changes hiring and evaluation. It changes education and healthcare. It changes finance and welfare. It changes administration and public opinion formation. AI policy is industrial policy, but it is also labor policy, education policy, and welfare policy. Furthermore, it is policy concerning the entire social system.
Prime Minister Kim Min-seok's recent reference to "AI for All" and "human-centered inclusive AI" should be read in this context. The point is not to view AI solely as a growth engine. The point is to view it as a social transition that changes the entirety of human life. The concern is that responsible use of AI, safe and trustworthy norms and standards, and control mechanisms for human safety and protection must be discussed together.
What is needed now, therefore, is social dialogue. But that dialogue cannot be left only to existing structures such as the Economic, Social and Labor Council. The changes brought by AI do not stop at labor issues. They are connected to consumer protection, personal information, copyright, fair competition, medical judgment, financial review, content trust, and national security. The participants in the discussion must also broaden. Rather than relying on the same organizations that always participate, the scope of participation must be expanded to include the voices of those who will actually be affected.
What matters is the starting point. AI is actually deployed, applied to employees, and brought into contact with consumers at companies. The AI that society experiences ultimately appears in corporate hiring systems, call centers, financial reviews, performance evaluations, content recommendations, and medical and welfare services. If so, the discussion of an AI social contract should also begin first on the corporate front line.
Such movements have already begun overseas. The Business Software Alliance (BSA), which includes global software companies such as Microsoft, Oracle, IBM, and SAP, released "Policy Solutions for Building Responsible AI" in 2024. It proposed risk-based approaches, corporate governance, high-risk AI impact assessments, high-risk AI testing, and protection of creators' rights to policymakers in each country. What is interesting is that this is not a government report but a proposal from the software industry. Companies were not trying to evade regulation. They proposed first the direction of AI rules that society could accept.
The case of Japan's Keidanren is also worth noting. Keidanren already proposed "AI Utilization Strategy: Toward the Realization of an AI-Ready Society" in 2019. In 2023, it expanded the discussion with "Realization of Society 5.0 for SDGs Through AI" and "AI Utilization Strategy II." The core is clear. AI companies cannot be the only ones to prepare. Companies, individuals, and social institutions must prepare together. Keidanren linked AI to Japan's Society 5.0, that is, a human-centered super-smart society. It viewed AI not merely as technology adoption but as a question of social design.
The lesson from these cases is clear. Business associations can set the agenda for social dialogue first. They do not have to wait for the government to provide the right answer. The industry should propose principles first, experts should verify them, and the political sphere and civil society should debate them. Based on this, the government should lead social consensus and institutionalize it. In a field that changes as quickly as AI, this approach is in fact more realistic. Regulation always lags behind technology. If so, social discussion must begin first at the sites where the technology is used.
AI competition is unavoidable. But the standard of competition cannot be speed alone. It is not the country that adopts AI fastest that survives. The countries that survive will be those that prevent technology from devouring human time, those with the capacity for social design. The same is true of corporate strategy. How quickly a company uses technology matters. But who first proposes an AI operating order that society can accept matters more. Only then can companies be sustainable.
AI can replace humans. But there is a more important fact. AI can also exhaust humans. When the burnout society meets AI, fatigue becomes more refined. It becomes quieter. What is needed now is not politics that blocks AI. It is social dialogue that places AI within a human order. The starting point cannot remain only with government committees. It must begin with the companies that understand the technology best, actually use it, and are the first to confirm its effects. Now is the time for companies to propose first the social contract of the AI era.
