A fingerprint identifies a person uniquely enough to convict them in court. So does a voice. The Korean National Police Agency and Cheil built their 2025 campaign around that second fact, which the phone scam industry had been banking on people not knowing, and turned voice recordings of actual scammers into the most unusual wanted posters in the history of law enforcement advertising.
The Scale of the Problem
Phone scam losses in South Korea nearly doubled in a single year. By 2025, financial losses from voice phishing exceeded one trillion Korean won, approximately $680 million USD. The tactics driving those losses had grown significantly more sophisticated alongside the numbers. Scammers were impersonating government officials, sending fake delivery alerts, and, increasingly, using AI-generated voice cloning to replicate the voices of family members. Elderly victims across South Korea were receiving calls from voices that sounded exactly like their grandchildren, their sons, their daughters, asking for emergency money.
The problem with conventional fraud prevention communication is that it focuses on the method: the script, the impersonation, the specific scenario. Methods change. Scripts evolve. The tactics that worked in 2022 were already obsolete by 2024. Every awareness campaign built around a specific method is already behind the criminals who have moved on.
The Insight Behind VOICE WANTED
The Korean National Police Agency and Cheil identified the one element of phone fraud that does not change when the script changes: the perpetrator’s voice. Just as fingerprints are unique to each individual, voices carry distinct characteristics in their vocal cords that remain consistent regardless of what the person is saying or what role they are playing. A scammer who impersonates a police officer today and a bank official tomorrow has a voiceprint that connects both calls to the same person.
That observation reoriented the entire investigative strategy. Rather than asking the public to recognise specific fraud scripts, VOICE WANTED asked them to record and report the voices themselves. The voice is the evidence that persists when every other method has changed.
The Wanted Posters
Cheil produced wanted posters using waveform graphs extracted from actual voiceprint samples of real phone scammers. The visual language borrowed directly from the conventional wanted poster format: a face, or in this case, the visual representation of a voice, accompanied by identifying information and a request for public assistance.
The waveform graphs were not decorative. They represented the actual acoustic signature of individual scammers, the same data that voice analysis software uses to match recordings against a database of known perpetrators. Each poster also included a virtual composite sketch derived from the voice data alongside the waveform. A QR code on each poster linked directly to an interactive platform where citizens could listen to real scam call examples, learn about the latest fraud tactics, and most critically, submit recordings of suspicious calls they had received.
How the Submission System Works
The campaign’s functional architecture is what separates VOICE WANTED from a conventional public awareness campaign. When a citizen submits a recording of a suspicious call through the platform, police extract the voiceprint from the recording and run it against an existing database of scammer voice files. A match confirms a suspect’s identity. A new voice adds to the database, strengthening the analysis model for future cases.
The system also creates a prevention layer beyond investigation. Voice samples collected from the database can be provided to telecommunications companies, enabling them to send automated warning alerts when a call from a suspected scammer’s voice pattern is detected before the recipient picks up. The campaign is simultaneously building a law enforcement tool and a real-time fraud prevention infrastructure.
The Creative Mechanism
The VOICE WANTED campaign works as advertising because it converts a passive audience into active participants without asking them to do anything they would not already do when they receive a suspicious call. Most people who receive a scam call hang up and move on. The campaign gives that moment a destination: report the call, submit the recording, contribute the voice to a database that makes the next person harder to deceive.
The wanted poster format gives the campaign its emotional register. Wanted posters are culturally universally understood as a call for public cooperation in identifying someone dangerous. By placing a voice waveform where a face would normally appear, Cheil makes the case that a voice is as identifying and as consequential as a photograph, and that the same civic duty that applies to a conventional criminal identification applies to a phone scammer whose face the public may never see.
The Broader Context
The campaign arrived at the precise moment when AI voice cloning had made the problem it was trying to solve significantly harder. If a scammer’s actual voice can be algorithmically altered to sound like a family member, the traditional advice to trust your ears is no longer sufficient. VOICE WANTED does not ask the public to trust their ears. It asks them to submit the audio and let forensic voice analysis do what human ears cannot.
South Korea’s phone scam crisis is not unique to the country. The combination of an ageing population, widespread smartphone adoption, and increasingly sophisticated AI tools for voice synthesis has produced similar patterns across Japan, the United States, Canada, and multiple European markets. VOICE WANTED is a locally developed response to a globally applicable problem, built around an investigative capability that scales precisely because it relies on the one thing that changes less than any fraud script: the voice of the person making the call.
Campaign Name: VOICE WANTED
Agency Name: Cheil Worldwide
Brand Name: Korean National Police Agency (KNPA)
Location: South Korea (national campaign; December 2025)
