KT is stepping up efforts to shape international standards for AI security as artificial intelligence becomes increasingly embedded in digital services and cybersecurity operations.
KT said September 21 that it participated in an international standards meeting of Study Group 17 (SG17) under the International Telecommunication Union’s Telecommunication Standardization Sector (ITU-T), held in Chongqing, China, from September 7 to 11.
The ITU-T SG17 meeting brings together cybersecurity experts from governments, companies and research institutions around the world to develop international standards covering cybersecurity, security architectures, privacy protection and AI security.
At the meeting, KT proposed a new international standards work item focused on guidelines for developing AI models specialized in cybersecurity using foundation models.
As generative AI and foundation models become more widely adopted, AI-based capabilities for detecting, analyzing and responding to cyber threats are also gaining importance. KT said standardized development principles are needed to ensure that cybersecurity-focused AI models can be developed and evaluated consistently.
The proposal submitted by KT includes recommendations covering cybersecurity knowledge sources and datasets for AI model development, technical considerations for building specialized security models, evaluation criteria throughout the model development lifecycle, and practices for continuous model improvement.
If adopted as an international standard, the guidelines could provide organizations with a common framework for developing and operating AI-powered cybersecurity solutions. KT expects this to help improve interoperability among security technologies while supporting more advanced threat detection, analysis and response.
KT also contributed to international standardization efforts surrounding digital trust.
ITU-T SG17 is developing a technical report aimed at defining the concepts of “trust” and “trustworthiness” in telecommunications and ICT environments and establishing the components and requirements of a Trust Framework that can be applied to real-world systems and services.
KT proposed revisions to the report, including improvements to key terminology and additional components and requirements for the Trust Framework. With support from the United Kingdom and India, the company helped bring the technical report to publication.
KT plans to continue work on linking the Trust Framework with cybersecurity reference architectures while expanding international discussions on digital trust in AI-enabled services.
The company’s participation in the meeting was led by Chang-oh Kim, Chief Privacy Officer and head of KT’s Privacy Protection Group, and Heung-yeol Yeom, chairman of KT’s Privacy Protection Advisory Committee.
Kim currently serves as chair of ITU-T SG17 Working Party 3, which focuses on trust, security strategies and cooperation. Yeom previously served as chair of ITU-T SG17 from 2016 to 2024 and continues to support KT’s international standardization activities.
KT plans to expand its work across AI security, privacy and digital trust, drawing on its experience in telecommunications, ICT, cybersecurity and privacy protection.
The company said it will continue collaborating with international experts and standards organizations to reflect Korean technology and industry experience in global information-security standards.
“As AI rapidly expands across industries, it is important for the international community to establish common standards for cybersecurity-focused AI models and trust frameworks for ICT services,” Yeom said. “We will continue working with KT to ensure that Korea’s technology and expertise are reflected in international standards for next-generation information security and privacy protection.”
Kim said cybersecurity, privacy and trust should no longer be treated as separate areas in the AI era.
“Security, privacy and trust need to be embedded across AI, data and services as a whole,” he said. “Building on KT’s expertise in telecommunications, ICT, cybersecurity and privacy protection, we will continue contributing to global standardization in AI Security, Privacy and Trust.”
