Leading scholars urged that artificial intelligence (AI) must be developed and deployed in ways that foster human growth to prevent a dystopian outcome. They argued that not only governments but also companies, educational institutions, and individuals must each play their part to avoid human subordination to AI.
Yoo Young-jin, professor of management at the London School of Economics, attended the "AI and Human Values Forum" at Seoul Forum 2026, recently held at the Shilla Hotel in Jung-gu, Seoul. "A more serious threat than AI replacing and surveilling human jobs is technology that does not allow humans to grow," Yoo said. "If we settle for a society where AI replaces what humans used to do and humans are merely compensated for it, the outcome will be the maximization of 'phantom GDP.'" Phantom GDP refers to an economic pattern in which AI drives economic growth but household income and ordinary citizens' lives fail to improve.
Yoo classified AI into two types from the perspective of "human growth." The first is "vacuum cleaner AI." "This type of AI replaces the labor humans used to perform, generating cost-saving effects," he said. "But the people replaced by AI do not become more capable and do not grow. Most organizations are introducing AI in this way."
The second type he presented is "scaffolding AI." "This is AI that does not replace what humans used to do but enables them to do things they previously could not," Yoo said, sharing the story of Paul Conyngham, a data scientist in Sydney, Australia. When Conyngham's pet dog Rosie was diagnosed with mast cell cancer, he used OpenAI's ChatGPT to predict treatment strategies and protein structures. Based on this, a nearby university produced a customized mRNA vaccine that cured the disease.
Just as scaffolding allows structures to be sturdily expanded at construction sites, humans should likewise expand their capabilities through AI, Yoo said. "When AI is used this way, one person's growth also brings utility to others," he said. "A 'generative externality' takes place."
Yoo proposed that governments, educational institutions, and companies should build AI utilization frameworks focused on enhancing human capabilities. "AI must be developed in a direction that enables humans to do what they previously could not," he said. "Nations need to establish principles that allow AI sovereignty to be realized at the individual level, and companies need to assign challenging tasks even to junior employees."
Shin Sang-kyu, professor at Ewha Institute for the Humanities at Ewha Womans University, who took the stage next, also called for "a paradigm shift in AI utilization." "Humans do not refer to a mere state of nature," Shin said. "Because technology determines who we are and what we can do, humans are a kind of 'cyborg-like being.'" He added, "That is why what matters is the manner of using technology. If we merely outsource thinking to AI and delegate decision-making, humans will become lazy."
"The world we now face is one in which a technological elite and Big Tech determine and govern the direction of the world's development," Shin said. "We need responses such as embedding democratic values in the AI development process and integrating ethical and political content into AI education."
What he particularly emphasized was "preparation for the possibility of structural unemployment." "If employment collapse and structural unemployment caused by AI are not impossible, the state and society must actively prepare for them," Shin said. "At the individual level, since working hours and paid labor are bound to decrease, people must now be able to define themselves in their daily lives."
Shin saw that Korea, paradoxically, has the potential to embrace such changes quickly. "When the Pew Research Center surveyed 17 countries on what makes life meaningful, Korea ranked 'material well-being' first, but on the contrary, occupation received only 6% of responses, the lowest," he said. "This proportion was the lowest among the 17 countries, which paradoxically means there is also room for Korea to adapt relatively easily to the 'post-labor transition.'"
In the panel discussion that followed, industry experts also expressed agreement on the importance of human values. Kwon Soon-il, chief strategy officer (CSO) of Upstage, said, "As someone who celebrates humanity, I believe AI will expand human possibilities, but the human sacrifices that may nevertheless arise must be addressed by quickly building social consensus. The later the consensus is formed, the greater the confusion will be." He added, "As people use AI, the tendency for humans to think for themselves is fading, so we need to discuss how to prevent this."
Kim Kyung-hoon, AI Safety leader at Kakao, also stressed, "Among the threats AI may pose, the most important is the problem of humans misusing AI. In particular, there is insufficient social consensus on how to handle cases where generative AI provides false information, so the government and companies must work together to create a public forum on this."
