South Korea's Ministry of Science and ICT is pushing to participate in OpenAI's Trusted Access Cybersecurity (TAC) programme.

Industry sources said on May 19 that the ministry plans to meet next week with OpenAI Chief Strategy Officer Jason Kwon (권) to flesh out the participating institution and the participation method for TAC. It is known to be discussing whether the participating entity should be the Korea Internet & Security Agency (KISA) or the AI Safety Institute, as well as how to proceed with the procedure.

A day earlier, the ministry and OpenAI held an "AI security working-level workshop" and formed a consensus on joining TAC. At the event, Sasha Baker (사샤 베이커), OpenAI's head of national security policy, visited South Korea to introduce the TAC programme and demonstrate key cybersecurity functions of the company's latest AI models.

Some domestic private companies are already participating in TAC. At an experts' meeting on May 8 to respond to global cybersecurity projects, Lee Jong-hyeok (이종혁), director of the Information Security Industry Division at the ministry, said, "It is confirmed that there are companies and institutions in Korea participating in TAC," adding, "It is difficult to disclose the specific number of participating companies and whether government agencies are included."

TAC participants are granted limited access to OpenAI's security-focused AI model, "GPT-5.5 Cyber". Global security firms such as CrowdStrike, Palo Alto Networks and Cloudflare are participating and linking and using OpenAI models in their security systems.