[Reading Science] The Real Face of AI Sweeping Nobel Prizes: Threatening Humanity and Improving Life?

Nobel Physics Winner Hilton and Chemistry Laureate Herschabis... AI Field Superstars
Known for AI Progress and Revolution with Deep Learning and AlphaGo
Once Colleagues, Now Completely Divergent Views
"Uncontrollable AI Smarter Than Humans"
"AI Accelerates Scientific Discovery"

The 2024 Nobel Prizes shook the foundations of science. The honor of the awards went not to humans with heads and hands, but to the development of artificial intelligence (AI) and discoveries made through AI.

Jeffrey Hinton, a renowned figure in the AI field and professor at the University of Toronto, received the Nobel Prize in Physics, while Demis Hassabis, CEO of Google DeepMind, was awarded the Nobel Prize in Chemistry, announcing to the world the scientific significance of AI technology.

Although the journeys of these two geniuses differ, it is clear that they played decisive roles in shaping the present and future of AI. Let us examine how they opened the AI era and what changes AI will bring to humanity's daily life.

In the AI field, Hinton and Hassabis can be respectively regarded as the ‘pioneer of deep learning’ and the ‘innovator of applied AI.’ Hinton introduced the concept of deep learning, breaking the ice and playing a crucial role in AI development, while Hassabis shocked the world by delivering the scene where AI won a Go match against a human through ‘AlphaGo,’ signaling the full-scale opening of the AI era. Their achievements demonstrated that AI can play an important role beyond mere technological innovation, extending into the realm of fundamental science. This was recorded as a historic moment officially recognizing that AI research can contribute to the expansion of human knowledge.

Hinton’s research revolutionized the learning methods of artificial neural networks, while Hassabis focused on applying theory to practical problem-solving. The Turing Award, often called the Nobel Prize of computer science, was jointly awarded to Hinton and the other three AI greats?Yann LeCun, Andrew Ng, and Yoshua Bengio?but the Nobel Prize spotlighted only Hinton, the starting point of the research. Hassabis, displaying his unique genius, surpassed Hinton’s prot?g?s, the four AI greats, to claim the Nobel Prize.

◇ AI Neural Networks and GPU Utilization: This Person’s Research = Hinton was born in 1947 in Wimbledon, England, and majored in psychology at Oxford University. Beyond psychology, he focused on how the human brain processes and stores information and devoted himself to artificial neural network research. In 1978, Hinton crossed the Atlantic to earn a PhD in cognitive science at the University of California, San Diego, laying the foundation for artificial neural network theory. At Carnegie Mellon University, he developed a new type of neural network called ‘Boltzmann machines.’ This research became an important foundation for deep learning. Boltzmann machines evolved from the ‘Hopfield network’ devised by John Hopfield, who shared the Nobel Prize in Physics with Hinton.

However, at that time, there was little expectation that neural network research would yield results like today’s. Securing research funding was also difficult. Eventually, Hinton left Carnegie Mellon University and moved to the University of Toronto in Canada, which had promised research support. This was a decisive moment that made the University of Toronto the global center of AI.

Then came a decisive paper on deep learning in 2006: ‘Deep Belief Networks.’ This model, built by stacking multiple layers of restricted Boltzmann machines, marked the beginning of the deep learning era. Through this research, Hinton demonstrated that neural networks with many layers could be effectively trained, opening a new chapter in AI research.

Baek Sang-yeop, former CEO of LG CNS, said, “The moment I saw Professor Hinton’s deep learning paper, I realized a new era of artificial intelligence had begun.”

Professor Hinton is also the pioneer of AI training using graphics processing units (GPUs). He confirmed that AI training via GPUs was far faster than using central processing units (CPUs) and encouraged fellow researchers to utilize GPUs, accelerating AI research.

In 2012, the appearance of AlexNet, an image recognition system developed by Hinton’s research team using NVIDIA’s GPUs, became a decisive milestone in AI development.

Hinton is not a genius type. He said, “I was surprised to hear about winning the Nobel Prize in Physics. I quit physics in my first year of university because I couldn’t handle complex math.”

◇ Chess Prodigy Opens the Path of AI Science = Compared to Hinton, Hassabis is a prodigy. Born in 1976 in London to a Greek father and a Chinese-Singaporean mother, Hassabis gained fame as a chess prodigy. He learned chess at age 8 and won the British Junior Chess Championship three years later. At 11, he also won adult tournaments. His genius revealed in chess extended into academia. He entered Cambridge University early at 16 and majored in computer science. At university, Hassabis focused on AI. His chess-based knowledge combined with computer science theory expanded his interest in AI applications.

Hassabis’s genius laid the foundation in the gaming field. In 1998, he joined Bullfrog Productions, a game development company, and at age 21 led the development of the game ‘Theme Park World.’ In 2001, he founded his own game development company, Elixir Studios.

Games were already a limited space for Hassabis’s dreams. He enrolled in a cognitive neuroscience PhD program at University College London (UCL) and began AI research.

In 2010, Hassabis co-founded ‘DeepMind’ with the goal of developing artificial general intelligence (AGI) and using it to solve the world’s most complex problems. Four years later, in 2014, Google acquired DeepMind, investing a whopping ?500 million (about $800 million). At the time, this was an enormous amount and marked the first big tech company to win a Nobel Prize. DeepMind became ‘Google DeepMind,’ and Hassabis led AI research.

It did not take long for Google DeepMind to astonish the world. In 2016, AlphaGo appeared. AlphaGo defeated Lee Sedol 9-dan, one of the world’s top Go players, in Seoul with a 4-1 victory. Contrary to expectations before the match, Lee Sedol won only once after a series of losses. Korea and the entire world were stunned by AI’s capabilities.

At that time, Hassabis dreamed another dream: scientific research. Professor Seok Cha-ok of Seoul National University said, “After Hassabis’s victory over Lee Sedol 9-dan in Seoul, he said he would focus on scientific research going forward.”

Hassabis put this into practice. He consecutively released AI for scientific fields such as AlphaFold and AlphaZero, siblings of AlphaGo. AlphaFold, in particular, attracted attention as AI studying proteins, the origin of life.

Professor Seok said, “The field that has received the most Nobel Prizes is biomolecular analysis. This time, it is surprising that computer calculations produced results comparable to experimental levels.”

The path created by Hassabis and DeepMind is sure to become a major thoroughfare. Numerous follow-up studies are expected to emerge, revealing the principles of nature, the origin of life, causes of diseases, and methods of healing.

◇ Collaboration, Separation, and Conflict = Hinton and Hassabis were colleagues at Google. Hassabis congratulated Hinton, who won the Nobel Prize a day earlier, referring to him as a former Google colleague.

Google acquired DNN Research, founded in 2013 by Hinton along with University of Toronto graduate students Ilya Sutskever and Alex Krizhevsky (both researchers of AlexNet), one year before acquiring DeepMind. Hassabis and Hinton worked together under the large umbrella of Google. They had different approaches. Although they cooperated toward the common goal of AI development, they are now heading in completely different directions.

In 2023, when OpenAI introduced ChatGPT and the excitement around generative AI surged, Hinton left Google and turned into a preacher warning of AI risks. He said AI could threaten humanity and expressed regret over nearly 50 years of work. He feared the technology he helped lay the foundation for might endanger humanity.

Hinton focused on emphasizing AI’s dangers in interviews arranged by the Nobel Committee and the University of Toronto commemorating his Nobel Prize. He seemed to use the Nobel Prize as an opportunity to alert the world to AI’s threats.

He said, “We can see benefits of AI in fields like healthcare, but it can also lead to bad outcomes.” Hinton stated, “Within 5 to 20 years, AI will surpass human intelligence, and no one knows if we can control it.”

Hinton did not hesitate to criticize Sam Altman, CEO of OpenAI, who is rapidly commercializing OpenAI. During the Nobel Prize press conference, he said, “I am especially proud that one of my students fired Sam Altman.” The student Hinton referred to was Ilya Sutskever, who led Altman’s ousting.

Hinton is also at odds with another prot?g?, Andrew Ng, over California’s proposed AI regulation bill. Ng opposes regulation, while Hinton actively supports it. Their heated exchanges are ongoing on X (formerly Twitter).

Hassabis still maintains a position of contributing to AI development. In a statement released after winning the Nobel Prize in Chemistry, Hassabis said, “I have devoted my career to advancing AI because I see its potential to improve billions of lives,” adding, “I hope AlphaFold will be remembered as the first example demonstrating AI’s amazing potential to accelerate scientific discovery.”