KT to develop AI tech to filter spam messages

South Korea's KT Corp. said on Tuesday it will work with the Korea Internet & Security Agency (KISA) to launch an artificial intelligence (AI) system next month to detect malicious URLs in text messages and block spam.

KT’s AI will identify phone numbers linked to spam and share the information with message relay operators in a blacklist format, the company said. KISA will verify the numbers.

KT’s AI clean messaging technology analyzes spam-related data and will flag URLs that lead to harmful websites or install smishing files.

The company also said a small large language model (sLLM) will distinguish illegal spam from legitimate messages by analyzing intent.

Write to Ji-Eun Jeong at jeong@hankyung.com