Korea Internet & Security Agency and KT Strengthen Spam Blocking Using AI Technology

KT 'AI Clean Messaging' Technology Analysis
Spam Phone Numbers Provided to KISA

The Korea Internet & Security Agency (KISA) and KT announced on the 24th that they have signed a business agreement at the KT Songpa office to strengthen spam blocking using artificial intelligence (AI) technology.

KT has signed a business agreement with KISA to eradicate illegal spam. (From left: Im Geon-ho, Executive Director of KT Communication Platform Business, Lee Young-ho, Executive Director of KT Fair Competition, Myung Je-hoon, Head of KT Enterprise Telecommunications Business Division, Kim Joo-young, Head of KISA Personal Information Safety Utilization Division, Jung Won-gi, Director of KISA Digital User Protection Group, Ahn Jung-eun, Team Leader of KISA Spam Policy Team)

KISA has been blocking mass texts sent from phone numbers on a blacklist by analyzing spam report data and sharing phone numbers that have been reported multiple times or redundantly with all text message relay companies to prevent harm to the public.

In addition, to facilitate the analysis of spam blacklists that frequently change the text of messages and to take the lead in spam blocking through checking for malicious website addresses and contextual analysis, KISA will cooperate with KT.

The main points of this agreement are: ▲(KISA) providing spam report data to KT and including spam numbers extracted using KT’s technology in the blacklist to be shared with all text message relay companies ▲(KT) providing spam phone numbers analyzed and extracted using the 'AI Clean Messaging' technology based on KISA’s spam report data to KISA. This will be implemented starting next month.

The AI Clean Messaging technology is a collection of AI models trained on various spam-related data. This system consists of a URL model that analyzes large amounts of data to find website addresses linking to malicious web pages or smishing installation files, and a lightweight large language model (sLLM) that distinguishes between messages with specific illegal spam intent and normal messages.

Myeong Je-hun, Head of KT Enterprise Communication Business Division, stated, "We will take the lead in actively utilizing the latest technologies such as AI to reduce domestic spam distribution and minimize user damage caused by illegal spam."

Kim Ju-young, Head of KISA’s Personal Information Safety Utilization Division, said, "We expect spam blocking to be further strengthened through this cooperation, and KISA will continue to expand collaboration with related organizations and further enhance its role as a spam response platform."