The Hellenic Police (EL.AS.) has dismantled a sophisticated criminal organization that allegedly utilized Artificial Intelligence (AI) tools to clone voices. According to official reports, the group's illicit profits are estimated to exceed €1 million, with their operations active since at least early July 2026.

The Anatomy of the Scam

The organization maintained a clear internal hierarchy, consisting of directors, "callers," and "collectors." The perpetrators initially posed as employees of the Hellenic Electricity Distribution Network Operator (DEDDIE), mobile providers, or fiber optic installation crews. If a potential victim expressed doubt, the scammers suggested confirming the procedure with a relative.

The group would then contact the relative, and during a brief conversation, they would capture a voice sample. Using AI tools, they replicated the relative's vocal characteristics to call the primary victim back. The familiar voice served as a powerful tool for building trust, leading victims to reveal where they stored cash and valuables or to allow the "technicians" into their homes.

Arrests and Legal Proceedings

The ring's operations were halted in Nea Ionia, Athens, following an operation by the Economic Crime Division. Two men, aged 23 and 25, were arrested shortly after stealing jewelry valued at approximately €30,000. Four additional individuals are included in the case file, three of whom have already been identified. Authorities also recovered a vehicle stolen in September, which was used for their operations.

Police Recommendations for Citizens

The Hellenic Police emphasize that voice recognition is no longer a sufficient method for identity verification. Citizens are advised to:

  • Immediately terminate any suspicious calls.
  • Directly call their relatives on a known phone number to verify claims.
  • Use "safety words" or personal questions known only to family members.
  • Never inform strangers about the presence of cash or jewelry in the house.
  • Always contact official company numbers to verify technical work requests.