Science Minister Bae Says DeepMind Partnership to Focus on AI Hub, Science Center, Safety

The Ministry of Science and ICT said its cooperation with Google DeepMind will move forward on three tracks: building a global AI hub, linking with the National Science AI Research Center, and conducting research on artificial intelligence safety. 

The ministry said Monday it signed a memorandum of understanding with DeepMind outlining the cooperation. Deputy Prime Minister and Science and ICT Minister Bae Kyung-hoon said a closed-door meeting also covered how the two sides would divide roles going forward. 

Bae said the government and industry plan to prepare follow-up discussions with DeepMind focused on the bio and medical sector, including securing data, improving model performance and setting strategies for broader adoption. 

The following is a Q&A with Bae.

- What were the key areas discussed in the closed-door meeting?
"Discussions centered on three tracks: cooperation on a global AI hub, linkage with the National Science AI Research Center, and AI safety research. Since DeepMind has strong interest in 'solving scientific problems' and 'AI safety,' we agreed to strengthen cooperation in those areas. We are still at an early stage after signing the MOU, and the specific division of roles will be finalized through further talks. The government and industry will prepare plans focused on the bio and medical sector — including securing data, advancing model performance and strategies for wider use — and continue discussions with DeepMind."

- What did you mean by mentioning 'Mythos' in your remarks?
"It was to stress the importance of an independent foundation model. Without our own model, it is difficult to secure competitiveness in security and cyber safety. There are areas where we can cooperate with DeepMind, but there are also areas we clearly need to build ourselves."

- Is there potential for cooperation between domestic companies' models and DeepMind?
"Rather than cooperation between individual models, we focused on the broader direction of 'AI-based scientific research' and 'AI safety.' We also shared that independent model development is underway in Korea."

- Have the participating universities been decided for Google's AI campus?
"Google has only announced an investment in an AI campus in Korea, and the participating universities have not been decided. However, we agreed on the need to expand people-to-people exchanges and discussed steps such as forming a working group."

- What was discussed about building AI guardrails?
"We agreed on the need for guardrails, in the sense that humans must be able to control AI even as it develops on its own. We also shared the view that guidelines and a regulatory framework are needed so all users — not just certain groups — can use AI safely."



* This article has been translated by AI.

Copyright ⓒ Aju Press All rights reserved.