TLDR: LG Uplus has formalized a strategic partnership with the Korean National Police Agency to significantly enhance efforts against voice phishing. This collaboration leverages LG Uplus’s advanced AI technology, including its ‘ixi-O’ call agent, for real-time detection, prevention, and tracking of fraudulent activities, aiming to safeguard public safety and financial well-being.
SEOUL, South Korea – In a significant move to bolster national security against financial fraud, LG Uplus has announced a comprehensive business agreement with the Korean National Police Agency to combat voice phishing through the strategic deployment of artificial intelligence (AI) technology. The partnership, formalized on August 22, 2025, at LG Uplus headquarters, aims to establish a robust, real-time response system against increasingly sophisticated voice phishing schemes.
The collaboration will see LG Uplus sharing critical data from its communication infrastructure and AI systems with the police. This includes detected abnormal patterns, logs from malicious app control servers, suspicious text message sender numbers, and information on relay devices. In return, the Korean National Police Agency will provide its extensive crime databases to LG Uplus, which were previously utilized to train LG Uplus’s AI call agent service, ‘ixi-O’. This reciprocal data exchange is designed to create a more intelligent and responsive defense mechanism against financial crimes.
The ‘ixi-O’ AI agent, a cornerstone of LG Uplus’s anti-phishing strategy, plays a crucial role in this initiative. The technology, which was unveiled at the MWC tech trade show in March and commercialized globally as an on-device AI, is capable of detecting AI-generated or manipulated voices during calls within five seconds, alerting users to potential fraud. It was trained on an extensive dataset of approximately 3,000 hours of call data, equivalent to about 2 million calls, employing voice activity detection, speech-to-text conversion, and anti-spoofing analysis to identify subtle, unnatural variations in pronunciations or high-frequency noise that human ears might miss. This on-device AI approach also prioritizes user privacy by not storing data externally.
Park Seong-joo, head of the National Investigation Headquarters, emphasized the severity of the issue, stating, ‘Voice phishing is a serious crime that infringes on public safety. This agreement will serve as a model case to prevent crimes through public and private collaboration.’ Hong Beom-sik, CEO of LG Uplus, echoed this sentiment, affirming the company’s commitment to its role as a responsible private entity in protecting citizens.
Beyond technological integration, the partnership extends to on-the-ground support. LG Uplus will designate its security specialty stores as ‘shelters in the field’ for victims of crime or individuals suspected of having malicious apps on their devices. An emergency response process will also be established, linking initial security consulting with direct handover to the police for immediate action. This comprehensive approach aims to strengthen the real-time response system from pre-awareness and immediate intervention to post-tracking of criminal activities.
Also Read:
- Colorado Bureau of Investigation Warns Public of Sophisticated Cybercrime Agent Impersonation Scam
- South Korea Reforms Venture Investment to Propel AI Startup Ecosystem
This initiative builds upon LG Uplus’s prior successes in combating voice phishing. Earlier efforts saw the company’s AI-powered damage prevention analysis system block over 5,000 malicious apps between February and April, preventing an estimated KRW209 billion (approximately $150.2 million) in fraud losses over just three months. The company has also partnered with the National Forensic Service to develop a system for detecting criminal voices by comparing caller voices in real-time against a database of known phishing criminals. These advancements are critical as AI-powered voice phishing attacks become increasingly sophisticated, with enterprises experiencing a significant rise in AI-related security breaches.
