KT Blocks Unwanted Spam Texts with AI... 99% Accuracy

AI Spam Blocking Service Launched... Free for KT Customers
Users Can Register Unwanted Numbers and Text Spam
IP-Based and Voice Phishing Blocking Services Also Planned

KT announced on the 25th that it will launch the 'AI Spam Reception Blocking Service,' which automatically blocks spam messages using artificial intelligence (AI) technology. This is a free additional service available to all KT customers.

A KT employee is analyzing spam messages using the artificial intelligence (AI) Clean Messaging system at the KT Customer Center in Gwanghwamun, Seoul. Photo by KT

The AI Spam Reception Blocking Service automatically blocks unwanted advertising spam messages using AI. In addition to the messages blocked by KT, users can register additional numbers or phrases they do not want to receive, and those numbers or messages will be automatically blocked. When a message is recognized as spam, it does not get classified into a spam folder but is completely prevented from reaching the user.

Unlike existing services where humans analyzed messages and updated databases, the AI system accurately identifies and removes spam messages, enabling the blocking of approximately 10 million additional spam messages annually. The accuracy of AI spam blocking is 99%, and the time required for spam processing has been reduced by half compared to before. KT has trained the system with over 1.5 million spam data entries daily through deep learning during the three years of service preparation.

KT announced plans to add various safety services this year, starting with this AI spam blocking service. In the first half of the year, it will establish an 'IP-based real-time spam blocking' system. Unlike the existing URL-based blocking method, which is difficult to block when spammers change URLs, this system tracks the IP that sent the URL to block spam.

KT AI Spam Blocking Platform/Photo by KT

Additionally, an emergency blocking system for voice phishing numbers called 'Circuit Breaker (tentative name)' will be introduced. This system immediately blocks numbers used for voice phishing for a certain period.

In the second half of the year, KT plans to launch a 'Spam Risk Level Display in Messages' service that combines AI and big data analysis technology to inform users of the spam risk level of messages, minimizing phishing risks and creating a safer messaging environment.

KT is working with the Seoul Metropolitan Police Agency to suspend suspicious lines suspected of voice phishing and smishing (personal information fraud via text messages). Starting from the 28th, KT plans to automatically extract 'bait numbers' from messages sent from suspended suspicious lines and suspend them additionally. Bait numbers refer to separate phone numbers written in the message body, different from the sender's number, used to bypass number blocking during phishing attempts. It is estimated that this system will be able to suspend about 300 to 500 bait numbers monthly. According to the Financial Supervisory Service, the average voice phishing damage per person last year reached 17 million KRW.

Furthermore, KT has formed the 'Company-wide Safety and Security First Achievement Task Force (TF)' to strengthen user protection and ensure customers can use safer communication services. On the 14th, KT was ranked first among telecommunications operators in the 'Excellent Business Operators for Autonomous Resolution of Consumer Complaints' selected by the Korea Consumer Agency.

Lee Byung-moo, Head of KT's Customer Experience Innovation Division, said, "We expanded the telecommunications fraud response council formed last year into the company-wide Safety and Security First Achievement TF. We will develop real-time blocking technology so that customers can use services with peace of mind and respond more effectively to spam issues."