LG Uplus announced on the 5th that it detected over 5,500 attempted voice phishing incidents in one month through its on-device AI-based "Anti-DeepVoice" technology. This technology uses artificial intelligence (AI) to identify manipulated voices and reduce the risk of voice phishing.
LG Uplus commercialized this technology through the "ixi-O" app at the end of June. As a result, it blocked approximately 5,500 phishing attempts in advance, preventing an average of 183 phishing incidents per day. According to the Korean National Police Agency, the average damage amount per voice phishing incident is about 53 million won, meaning this technology is estimated to have prevented damages worth around 290 billion won.
"Anti-DeepVoice" detects manipulated voices within 5 seconds after a call and determines whether a voice phishing attempt is suspicious within 1 to 2 minutes. This technology detects synthetic voices with 98% accuracy and provides real-time protection without personal data leakage using on-device AI.
This technology has garnered attention in the global market, with the industry security chief of the Global System for Mobile Communications Association (GSMA) evaluating "Anti-DeepVoice" as the new benchmark for mobile security. LG Uplus plans to enhance customer security further by adding voice detection for voice phishing criminals and AI-based detection features for risky URLs and malicious apps in the future.