AI does not replace human work. It will instead create more and new kinds of work.
Fei-Fei Li, dubbed the "godmother of AI," a professor of computer science at Stanford University and co-founder of the Institute for Human-Centered AI (HAI), said this at the Seoul Digital Forum (SDF) 2025 held at Dongdaemun Design Plaza (DDP) in Seoul on the 13th. She added, "A hundred years ago, 80% of humanity were farmers, but now it's under 10%," noting, "Work has not disappeared; it has evolved."
Li gave a lecture on the theme "The direction of AI technology's evolution and its social meaning." Li emphasized that "AI is a tool that complements human capabilities, and design that puts people before technology must be the core of future AI."
Li cited changes in education as the most important task. "Teaching is a job that will never disappear," Li said, adding, "Humans are learners, and younger generations inevitably learn from adults." Li added, "AI is not a technology that replaces teachers but a means to expand learning and creativity, and education must shift from memorization to cultivating creativity and problem-solving skills."
Li also pointed out the concentration of AI resources. "Today's AI industry concentrates funding and data in a few corporations and countries," Li said, adding, "To mitigate this, government and society must actively encourage the open-source ecosystem and entrepreneurship." Li emphasized, "A sustainable AI ecosystem can be built only when not only large corporations but also Start - Up, academia, public institutions, and nonprofits participate together."
Li also addressed the AI anxiety spreading across society. "Much of the anxiety about AI stems from a lack of information," Li said, adding, "Governments and educational institutions must clearly explain to citizens the principles and limits of AI and strengthen public education." Li added, "As AI affects democracy, labor, and the overall social structure, citizens must be able to understand and control the technology."
Assessing ChatGPT, called a turning point for Generative AI, as "a watershed moment in the popularization of AI," Li argued that AI's next turning points are "spatial intelligence" and "large world models." Spatial intelligence refers to AI's ability to understand relationships between objects in three-dimensional (3D) space and to think, move, and create in real environments. Large world models are technologies that learn such spatial information to autonomously construct and generate virtual worlds, which can be applied to fields such as robotics, Autonomous Driving, and digital fabrication.
"Language is a one-dimensional signal, but the real world is three-dimensional—four-dimensional when time is included," Li said, predicting, "AI that can understand the world spatially and recreate it will be the core of future innovation." Li explained, "This technology is likely to be applied across industries that interact with physical space, including robotics, manufacturing, architecture, and art."
Li took a cautious stance on artificial superintelligence (ASI). "AI already surpasses humans in specific tasks like translation or image generation, but it has not yet reached human-specific intelligences such as emotion, empathy, and sociality," Li said. Li added, "Human reasoning that observed apples falling and discovered the law of gravity remains an area AI has yet to catch up with."
Li also gave high marks to Korea's AI industry. "Korea is a globally passionate country with a high level of technology," Li said, adding, "It has secured global competitiveness based on excellent STEM (science, technology, engineering, and mathematics) education and research infrastructure." Li added, "Not only large corporations like Samsung but also the active challenges of Start - Up are impressive."
"AI is a tool, and its direction is decided by humans, and for AI to be used for good, every citizen must participate," Li said. "Not only engineers but also artists, teachers, doctors, students, and citizens must all be the protagonists of the AI era."