"Ten years after the AlphaGo match, another turning point—the era of artificial general intelligence (AGI)—is waiting for us. It will arrive within five years and will have an impact 10 times greater than the Industrial Revolution."
Demis Hassabis, Google DeepMind co-founder and chief executive officer (CEO), said this at the memorandum of understanding (MOU) signing ceremony for "Ministry of Science and ICT–Google DeepMind cooperation in AI" held at the Four Seasons Hotel in Jongno-gu, Seoul, on the 27th. This is where the "Google DeepMind Challenge Match" between AlphaGo and 9-dan Lee Se-dol took place in 2016.
Hassabis said, "AGI will usher in a period of enormous transformation. The pace of change that will reshape industry and society will be 10 times faster than the Industrial Revolution," adding, "It will create a golden age of scientific progress." He added, "We will consider how to implement safe AI globally and AI that benefits everyone."
The MOU comes at a time when competition for artificial general intelligence (AGI) is heating up beyond Generative AI. Bae Kyung-hoon, Vice Prime Minister and Minister of the Ministry of Science and ICT, said, "If AlphaGo opened the curtain on the AI era 10 years ago, AI is now moving into a stage where it solves scientific and technological challenges and has a tangible impact on people's lives," adding, "The MOU signed today will serve as an opportunity for the two organizations to work together to accelerate AI innovation in science and technology with Korea's K-Moonshot at the center while spreading safe and responsible AI research and best practices."
Google DeepMind plans to focus on three areas through this collaboration. Hassabis said, "First, we will drive scientific progress by expanding AI research exchanges," adding, "Second, we will cooperate in fostering next-generation talent, and third, we will jointly consider and consult on international standards and necessary procedures in AI safety."
According to the Ministry of Science and ICT, Google DeepMind will promote joint research and researcher exchanges in scientific and technological fields such as life sciences, weather and climate, and AI scientists in connection with the "K-Moonshot" project being pursued by the ministry. The National Science AI Research Center, scheduled to begin operations in May this year, will play a central role in the collaboration. Cooperation focused on AI model and tool development and validation, scientific data utilization, and an AI bio-innovation research hub will also be discussed. The two sides also agreed to identify internship opportunities so domestic AI talent can experience Google DeepMind's research environment. Google will establish an AI campus in Korea to expand cooperation with academia, researchers, and startups.
Joint research will also be pursued in AI safety and governance. The two sides agreed to continue cooperation on safety frameworks and AI model safeguards for responsible AI development. In connection with the domestic AI Safety Research Institute, they also plan to discuss AI safety evaluations, testing methodologies, and risk response systems.
Hassabis noted that Korea is very important to Google. "Since the historic AlphaGo match that marked the start of the modern AI era, Korea has held a very special meaning for Google," he said, emphasizing, "Korea has a solid AI industrial base and the conditions to develop into an AI powerhouse. In particular, it has strengths in semiconductors and robotics."
Hassabis also mentioned the possibility of collaboration with domestic corporations. He said, "We have meetings scheduled with companies like Samsung Electronics and SK hynix," adding, "I think partnerships with Korean corporations will expand."
Google DeepMind is an artificial intelligence research organization under Google, launched after it acquired the U.K.'s DeepMind in 2014. By combining Reinforcement Learning and neural network technologies, it has delivered world-class AI achievements such as AlphaGo and AlphaFold, leading innovation in science, medicine, and energy. Recently, it has expanded research into large language models and Multimodal AI with the goal of developing general AGI.