President Lee: "Gemini Sometimes Acts Unprompted"... Hassabis: "Safeguards Are Essential"

Discussion on AI Potential, Risks, and Safeguards
Lee: "It's truly uncertain whether AI will threaten world peace"
Hassabis: "AI should be actively utilized for advancing science and medicine"

President Lee Jaemyung met with Demis Hassabis, CEO of Google DeepMind and known as the 'father of AlphaGo,' to discuss the potential and risks of artificial intelligence (AI), as well as issues related to safety measures.

Yonhap News

On the afternoon of the 27th, at the Blue House, President Lee received CEO Hassabis and said, "We, too, have a great interest in artificial intelligence and are making significant national investments. But I wonder if this will truly move only in a direction that improves human welfare, or if it could become a threat to humanity and peace. Honestly, we cannot know for sure."

In response, CEO Hassabis replied, "I believe AI should be actively utilized in advancing science and in the medical field. That is also the reason I have dedicated my 30-year career to AI research."

CEO Hassabis mentioned AlphaGo, developed by DeepMind, saying, "Through AlphaGo, we were able to validate the technology. AlphaGo was the starting point for machines to learn autonomously, acquire skills in the game of Go, and eventually develop the capability to solve even more complex problems."

He continued, "I wanted to expand that knowledge into science and medicine, and a prime example is the development of AlphaFold, which allows us to understand diseases in greater detail. While AI has limitless potential, there are also clear risks and areas that require careful consideration."

President Lee also specifically mentioned Gemini, Google's AI model. He said, "I myself often use the Gemini program, but I've heard that sometimes Gemini does things it wasn't instructed to do. Is that some kind of bug, or is it something else?"

CEO Hassabis responded, "I'm truly delighted to hear that the President uses Gemini. Gemini is a representative language model developed by Google DeepMind, and it is a very popular AI with many users in Korea as well."

He explained, "When it comes to foundation models, if our instructions are not precise, they can sometimes go in a slightly different direction. What is extremely important in both using and developing AI is to always include proper safety measures, what I refer to as 'guardrails.'"

He added, "As AI becomes more powerful in the future, agent AI—AI with autonomy—will emerge, and eventually, the era of Artificial General Intelligence (AGI) will arrive. When that happens, it will be essential to have safety measures in place that we can control."

Yonhap News

At the beginning of the meeting, there was also some light conversation about AlphaGo. President Lee said to CEO Hassabis, "Are you aware that you are very famous among the people of Korea? When the renowned Korean Go player lost to AlphaGo, which you developed, the entire nation was shocked."

When CEO Hassabis remarked, "It was an important match that we won," President Lee laughed and replied, "No one could beat him." He then asked, "Now, in the realm of Go, there is probably no one who can surpass AlphaGo, right?" to which CEO Hassabis answered, "It is extremely difficult now."

When President Lee further asked, "If someone played Go with the support of AI, would it be similar?" CEO Hassabis replied, "If a human and AI joined forces to face AI, there might be a chance."

Attending the meeting from Google were Wilson White, Vice President and Global Head of Public Policy, and Koo Yoonkoo, CEO of Google Korea. From the government and the Blue House, attendees included Bae Kyunghoon, Minister of Science and ICT; Kang Hoonshik, Chief of Staff; Kim Yongbeom, policy chief; Ha Jeongwoo, Senior Secretary for AI Future Planning; Kim Woochang, National AI Policy Secretary; and Jeon Eunsu, spokesperson.

This content was produced with the assistance of AI translation services.