OpenAI said on Thursday it held a "Cyber CXO Roundtable" at the Westin Seoul Parnas, inviting about 60 executives including chief information security officers at South Korean companies.

At the event, Kim Kyung-hoon (김경훈), head of OpenAI Korea, along with Dominic Grillo, head of technical success, and Wana Tun, a security specialist engineer, introduced AI-based strategies for responding to cyber threats.

OpenAI said companies are seeing shorter security response times as AI improves its ability to detect vulnerabilities and analyse attack paths. It stressed the importance of proactive defence systems that find and fix problems before attackers exploit vulnerabilities.

OpenAI also introduced its AI-based integrated cyber defence solution, Daybreak. Daybreak links OpenAI's AI models with Codex and Codex Security to support everything from vulnerability detection to verifying exploitability, preparing fixes and confirming their effectiveness.

OpenAI shared its experience building Defense Factory, its own operating model that connects AI agents with existing security and development tools to continually check for and address vulnerabilities.

It particularly stressed that companies should clearly define the data, systems and task scope that AI can access, and that people must review and approve key security measures.

Kim said companies need to treat cybersecurity as a core management task to quickly identify and respond to major security risks in the AI era. He said OpenAI would help South Korean companies apply AI to existing security systems, confirm its effectiveness on the ground and develop it into sustained defensive capabilities.