Vice Prime Minister and Minister Bae Kyung-hoon of the Ministry of Science and ICT said on the 16th that to respond to cyber threats using artificial intelligence (AI), the government will release a security-specialized proprietary AI model within the year.
Bae said at a presidential policy briefing held at the Blue House that day, "We are pushing within the year to create a security-specialized model by further training our existing proprietary AI model with security-related data," adding, "AI can be used to easily find and exploit security vulnerabilities, so preparation is needed on the defense side."
Bae added that while the government will build a security-specialized model within the year by further training the existing proprietary model with security data, in the mid to long term it will consider developing an advanced model on par with Anthropic Mythos. Bae said, "An advanced frontier model like Mythos can easily solve security problems even though it is not a model targeting only security," adding, "We need to consider developing an advanced frontier model."
Bae said Korea's AI competitiveness was evaluated as third in the world by an international evaluation agency, and that the government will seek to break into the second-place tier globally. Bae said, "When the second-round AI model evaluation results come out in Aug., we expect we may be able to go beyond third place and challenge for the second-place tier."
The uncertainty surrounding U.S. controls on AI access was also discussed at the briefing. President Lee Jae-myung noted, "We must prepare on the assumption that access will be blocked," adding, "There must not be a situation where another country is guarding our security and then suddenly opens the door." Bae said, "The United States banned AI access, then lifted it, and recently has been controlling it again in a limited way," adding, "China shows similar signs, so we do not know when it might be blocked."