LG Uplus announced on July 16 that its advanced artificial intelligence (AI)-driven spam blocking system resulted in a 1.4-times increase in blocked spam messages over approximately five months from February to June of this year. As scammer tactics such as voice phishing and smishing become more sophisticated, tailored AI filtering has proven to be effective in combatting these threats.
These deceptive messages, known as ‘bait texts,’ often impersonate government agencies, financial institutions, or courier services, or disguise themselves as family members to lure recipients into clicking malicious URLs or making financial transactions. Recently, scammers have exploited high-interest keywords like ‘government support coupons’ to induce users to download malicious apps or engage in other harmful actions.
The challenge lies in the rise of spam messages containing content very similar to legitimate texts, making simple keyword-based blocking methods problematic, as they may inadvertently block genuine messages. Conversely, loosening filtering thresholds increases the risk of exposing users to actual spam, potentially leading to customer damage.
To address this, LG Uplus has upgraded its ‘Customer Damage Prevention Analysis System,’ integrating real-time detection and blocking of suspicious spam messages. The system analyzes multiple tactics used by scammers, such as mimicked phrasing and manipulated caller IDs, to respond effectively.
Initially, the AI continuously learns from accumulated spam data and current messaging trends, updating the patterns of messages to block in real-time. It also assesses sender legitimacy, including verification of whether a message was sent from overseas or if the sender’s number has been manipulated. For example, even if a message contains words like ‘card delivery,’ if the sender’s number appears tampered with or originates from abroad, the message is automatically classified as spam.
Since its full deployment in February, the system has demonstrated impressive results. By June, the number of blocked spam messages increased by approximately 1.4 times compared to January, and customer reports of spam texts decreased by about 65%. The rate of real-time automatic detection and blocking by AI also nearly increased eightfold. LG Uplus states that these refinements significantly reduce customer inconveniences caused by spam messages.
Looking ahead, LG Uplus plans to further enhance its spam filtering system in the second half of the year. New features will be introduced to detect spam pattern changes more swiftly, and additional technological improvements will be undertaken to boost AI prediction accuracy.
Hong Kwan-hee, Executive Director of LG Uplus’s Security Center, said, “Countering increasingly sophisticated spam methods with advanced technology is crucial to protecting our customers.”
