TLDRs:
- LG Uplus is launching a real-time AI tool to detect fake voices during phone calls, targeting deepfake and phishing scams.
- The system flags manipulated voices in just five seconds without sending user data to external servers.
- The tool works directly on devices to improve privacy and integrates with national law enforcement voice databases.
- LG Uplus is repositioning itself as a leader in telecom security amid a global rise in AI-enabled scams.
South Korea’s LG Uplus is taking a bold step in the fight against AI-driven phone scams with the upcoming release of a real-time voice authentication tool that detects fake or manipulated voices during calls.
Set to launch on June 30 as part of its ixi-O AI agent, the feature marks a significant advancement in the telecom industry’s response to deepfake-enabled phishing threats.
AI Voice Detection in Real Time
The system has been trained on over 3,000 hours of voice data, covering approximately 2 million calls. With that foundation, it can identify subtle abnormalities in voice patterns, such as unnatural pronunciation or frequency distortions, often present in AI-generated voices. Within just five seconds of a call starting, users receive alerts if the system detects any signs of fraud, including impersonation or fabricated emergency situations.
This move comes amid growing concerns over voice phishing, where scammers use advanced voice-cloning technology to deceive victims by pretending to be relatives, colleagues, or even government officials.
These schemes have become more convincing as artificial intelligence continues to evolve, pushing telecom providers like LG Uplus to rethink their role in user protection.
Privacy-First Approach with On-Device AI
Unlike traditional server-based approaches, this feature is designed to operate directly on the user’s device. That decision reflects a broader trend in AI development toward protecting privacy by avoiding the external storage or transmission of sensitive user data. By keeping voice analysis local, LG Uplus reduces the risk of information leakage while still offering sophisticated fraud detection.
The anti-phishing tool also includes a backend layer of protection. LG Uplus plans to analyze calling patterns from flagged numbers and cross-reference them with a national criminal voice database maintained by the National Forensic Service. This collaboration allows authorities to trace and link fraudulent activity across different incidents, enhancing the broader cybersecurity landscape in South Korea.
Rising Threats Prompt Industry Action
The timing of this rollout is critical. Voice phishing scams have surged globally as AI voice synthesis tools become more accessible. Industry data shows that AI-related security breaches are not only increasing in frequency but are also harder to detect.
Compared to traditional attacks, AI-powered incidents take significantly longer to identify and contain, with major financial and reputational consequences.
LG Uplus’s broader strategy also hints at a shift in how telecom companies position themselves in the digital era. By branding itself as the security-first carrier, the company is moving beyond its traditional role as a network provider. Its efforts now include proactive threat prevention, real-time AI monitoring, and institutional partnerships to support national security efforts.
Telecoms Transform into Security Guardians
Notably, this is not the first time LG Uplus has deployed artificial intelligence in the fight against fraud. Earlier this year, it upgraded its damage prevention systems using AI, reportedly helping avert over $150 million in potential financial losses. With this latest move, the company continues to push the boundaries of what telecommunications security can offer in an AI-driven world.
