Cho Sung-kap, Ph.D.
President, Asia America AI Tech. University
Sir Isaac Newton (1642–1727), who formulated the law of universal gravitation, is said to have watched an apple fall and wondered why objects fall toward the ground. His persistent pursuit of that deceptively simple question ultimately led to one of the foundational laws of modern physics. Newton remains one of history’s greatest scientists, having helped lay the foundations of both physics and mathematics.
Born prematurely in Woolsthorpe, Lincolnshire, England, on Christmas Day in 1642, Newton lost his father before he was born. He entered Trinity College at the University of Cambridge, where he initially studied classical philosophy. He soon became absorbed in the scientific and mathematical ideas of René Descartes, Galileo Galilei and Johannes Kepler. Through relentless and original inquiry, he went on to change the course of scientific history.
Korea also has a distinguished history of scientific and technological achievement. Goryeo celadon, Silla’s textile craftsmanship, Jang Yeong-sil’s rain gauge and Choe Mu-seon’s gunpowder technology all demonstrate the sophistication of Korean innovation in their respective eras. The Jikji, printed in 1377, preceded Gutenberg’s metal-type Bible in Europe by approximately 78 years. Admiral Yi Sun-sin’s turtle ship is another remarkable achievement that holds an enduring place in world history. These accomplishments remain a profound source of pride for later generations.
Yet in the age of artificial intelligence, which has emerged on the shoulders of such giants, Korea is racing to adopt large language models developed overseas. Watching this unfold leaves me with a deep sense of responsibility. Why does Korea still struggle to produce individuals capable of changing the direction of global science and technology?
Behind today’s generative AI systems and the ability to predict protein structures with extraordinary speed stand researchers once dismissed for pursuing ideas considered impractical or impossible. These “alchemists of AI” stayed with their work when few believed in its value. The lives and achievements of seven pioneers who helped shape the AI era offer a valuable lens through which to examine the challenges facing Korea.
Seven Pioneers Who Shaped the AI Era
① Geoffrey Hinton — The Godfather of AI
Geoffrey Hinton is a British-Canadian computer scientist born in England in 1947. After earning a Ph.D. in artificial intelligence from the University of Edinburgh, he continued to pursue neural network research based on the conviction that computers could learn in ways resembling the human brain.
During the “AI winter” of the 1980s and 1990s, research funding disappeared and neural networks fell out of favor. Hinton was often treated as an outsider, yet he continued his work at the University of Toronto. He made major contributions to the development and wider adoption of backpropagation and helped ignite the deep-learning revolution through AlexNet, the image-recognition system unveiled in 2012. After receiving the Turing Award in 2018, he was awarded the Nobel Prize in Physics in 2024.
② Yann LeCun — A Pioneer of Visual Intelligence
Born in France and long based in the United States, Yann LeCun earned his doctorate from Pierre and Marie Curie University in Paris. He later worked at AT&T Bell Laboratories and New York University and now serves as Meta’s chief AI scientist.
At a time when much of the scientific community had turned away from neural networks, LeCun pursued architectures designed specifically for image recognition. His best-known contribution is the convolutional neural network, or CNN.
Traditional fully connected neural networks processed an image by flattening its many pixels into a one-dimensional stream of data. This approach weakened spatial information and caused the number of learning parameters to grow rapidly. CNNs addressed these limitations by extracting significant features while preserving the image’s two- or three-dimensional grid structure.
CNNs are particularly effective at processing images and spatial information, while recurrent neural networks, or RNNs, are designed for sequential data such as language and time series. LeCun’s research supported handwriting-recognition systems used by American banks in the 1990s. It later became an essential foundation for autonomous driving, facial recognition and AI-based medical imaging.
③ Demis Hassabis — From Chess Prodigy to Nobel Laureate
British-born Demis Hassabis reached master-level strength in chess by the age of 13. After studying computer science at the University of Cambridge, he worked as a video-game developer and produced several successful titles.
Convinced that building AI required an understanding of how the human brain works, Hassabis later earned a Ph.D. in cognitive neuroscience from University College London. He co-founded DeepMind in 2010.
Algorithms initially explored through games eventually became important components of modern AI. In 2016, AlphaGo stunned both the international Go community and the wider scientific world. DeepMind later developed AlphaFold, an AI system that transformed protein-structure prediction and reshaped research in biology and drug discovery. Hassabis received the Nobel Prize in Chemistry in 2024 for this achievement.
④ Timnit Gebru — A Pioneer of Responsible AI
Timnit Gebru was born in Addis Ababa, Ethiopia, in 1983. At the age of 15, she moved to the United States after fleeing the Ethiopian-Eritrean War.
Before entering Stanford University’s doctoral program, Gebru worked at Apple on signal-processing and computer-vision technologies. She later co-authored the influential Gender Shades study, which demonstrated that commercial facial-analysis systems produced significantly higher error rates when identifying women and people with darker skin tones.
After serving as co-lead of Google’s Ethical AI team, Gebru founded the Distributed Artificial Intelligence Research Institute, or DAIR. She continues to advance a more inclusive, equitable and accountable approach to AI development.
⑤ Anima Anandkumar — Using AI to Model the Natural World
Anima Anandkumar was born in Mysore, India. Beginning her education in India, she went on to become one of the world’s leading AI researchers.
Her work has advanced neural operators, a class of AI methods capable of rapidly simulating complex natural phenomena such as physical processes, weather patterns, fluid dynamics and fusion reactions.
As a professor at the California Institute of Technology and a former senior director of AI research at NVIDIA, Anandkumar has played a central role in applying AI to drug discovery, climate modeling, semiconductor design and other scientific fields.
⑥ Ali Farhadi — Advancing Real-Time Object Recognition
Ali Farhadi was born in Tehran, Iran, and completed his early education there before moving to the United States. He earned a Ph.D. in computer science from the University of Illinois Urbana-Champaign.
Farhadi contributed to the development of You Only Look Once, better known as YOLO, a real-time object-detection algorithm. YOLO transformed computer vision by identifying the locations and categories of multiple objects within an image in a single processing pass. It has become an important technology for autonomous vehicles, video surveillance and robotic vision.
Farhadi now serves as chief executive of the Allen Institute for AI, or AI2. The institute contributes to an open AI ecosystem by releasing models such as OLMo, along with research findings and development resources, under open-source frameworks.
⑦ Andrew Ng — Lowering the Barriers to AI Education
Andrew Ng was born in the United Kingdom and grew up in Hong Kong and Singapore. He has played a major role in popularizing deep learning and making AI education accessible to a global audience.
Ng co-founded the Google Brain project and later led AI research at Baidu. He also co-founded Coursera, one of the world’s largest online learning platforms, enabling millions of people to study artificial intelligence and computer science.
His contribution extends beyond the development of new algorithms. He helped turn AI from a specialized field accessible mainly to elite researchers and institutions into a body of knowledge that people around the world could study.
Why Does Korea Struggle to Produce Such Pioneers?
Throughout its long history, Korea has demonstrated remarkable creativity and technological capability. Why, then, does it remain difficult for the country to produce researchers capable of reshaping the direction of global science and technology?
Korea has a strong ambition to lead in advanced technologies. Yet systemic differences make it difficult for figures such as Geoffrey Hinton or Demis Hassabis to emerge and flourish in the Korean research environment. A culture unwilling to wait for unconventional research, an education system that does not adequately nurture creativity, and traditional values that place excessive emphasis on status and prestige continue to constrain the development of original talent. I write this because Korea must leave a happier and more prosperous country to the next generation.
Hinton was able to devote more than three decades to neural network research despite repeated claims that his work had little practical value. His persistence was made possible by a basic-science ecosystem willing to support research even when immediate results were unlikely.
By contrast, Korean researchers are frequently expected to produce visible outcomes within one to three years. Publication counts and near-term commercialization results often dominate performance assessments. Under such conditions, basic or unconventional research requiring years of uncertainty and repeated failure has little room to survive.
Korea also lacks sufficient flexibility for researchers to cross disciplinary boundaries. Hassabis moved through the worlds of chess, video-game development, neuroscience and AI, dismantling traditional academic divisions along the way. Korean education and research institutions, however, remain constrained by rigid career tracks, intense concentration of talent in medical schools and institutional barriers between academic departments. Such structures leave limited space for interdisciplinary innovators to grow.
An Ecosystem That Attracts Global Talent and Capital
There is also a substantial gap in the ability to attract global talent and capital. As discussed in the document How the United States Became the World’s Most Powerful Nation, America’s strength lies partly in its capacity to draw outstanding people from around the world and turn their knowledge into national scientific and industrial assets.
The capital of Big Tech companies such as Google and Meta, combined with vast computing infrastructure, enables researchers to pursue unconventional and disruptive ideas. A culture that recognizes ability and achievement regardless of nationality also draws researchers from across the world to the United States.
Korea, meanwhile, continues to worry about the loss of its best minds to other countries. It also faces a considerable gap in the capital and supercomputing infrastructure required to train large-scale AI models. The country must therefore do more than educate talented people at home. It must build an environment in which world-class researchers actively choose to come, work and remain.
Producing leaders for the AI age requires far more than training engineers who can write good code. Korea needs a long-term research environment in which scholars can survive while pursuing unconventional ideas for decades. Failure must be accepted as part of inquiry, and researchers must be free to connect disciplines that have traditionally remained separate.
Korea must also create an open ecosystem capable of attracting talent from around the world. Its future as an AI power will not be determined solely by how quickly it adopts models developed overseas. It will depend on how patiently the country supports people who pursue problems without established answers and how long it is willing to stand behind those who seek to open paths no one has traveled before.
