A familiar voice no longer proves who is calling
Voice-cloning software can imitate aspects of a real person's speech after obtaining recordings online. A scammer may claim to be a family member in an accident, a manager who needs a rushed transfer or a representative of a trusted institution. Latvia's CERT.LV has documented a local case in which a caller imitated a friend and demanded urgent money; an independent callback exposed the deception. Ordinary caller-ID information can also be spoofed. Treat a convincing voice as one clue, never as sufficient identity verification.
Pause when the caller demands secrecy or speed
A classic script combines emotional pressure with an unusual payment request: an accident, arrest, damaged phone, emergency medical bill or supposedly confidential company transfer. The caller may say there is no time to speak to anyone else. That urgency is a reason to slow down, not to obey. Do not give banking credentials, Smart-ID approvals, one-time codes, card details or account access over the phone. If the request involves funds or sensitive documents, stop the conversation and verify independently before touching your banking application.
Call back through a trusted route
Hang up and dial the person's number from your own contacts, not a number supplied during the suspicious call. If they cannot answer, contact another family member or colleague through a separate channel. For a bank, public agency or workplace request, use the published phone number or an internal directory. A second message from the same compromised social account is not necessarily independent confirmation. The important test is whether a trusted communication path, separate from the urgent caller, confirms the story. A caller's refusal to allow verification is a major warning.
Agree on a family verification plan in advance
Families can discuss how they would ask for help if a phone is lost or someone is travelling. Set a simple procedure: call a stored number, ask another trusted person to confirm, and never send money solely on an unexpected call. A shared phrase can help in some situations, but it should not be the only defence, especially if it could be overheard, guessed or revealed online. For companies, require a second approval for changes to bank details or urgent transfers, ideally using a different channel and a known contact.
Do not trust a video solely because it looks realistic
AI-generated or manipulated videos can place a familiar face into a fake advertisement or video call. Visual glitches, odd speech rhythms or inconsistent lighting may raise suspicion, but modern fakes can look natural, and genuine recordings can have similar artefacts. Instead of relying on a checklist of pixel errors, inspect the source, find the original publication and verify the claim through the subject's trusted official channel. CERT.LV has described Latvian cases of stolen identities and deepfake endorsements; the risk is real even when the video is shared by someone you know.
If you already sent money or revealed a code
Contact your bank or payment provider immediately, explain that the transaction may be fraudulent and ask about blocking, recalls or disputing it. Change affected passwords from a trusted device, review sessions and contact the organisation whose identity was impersonated. Preserve the phone number, messages, times, screenshots and payment records. Report the incident to relevant local authorities or national cybersecurity channels. Do not hire a stranger promising guaranteed recovery. The safest habit is not learning to identify every synthetic voice: it is refusing to act until a separate trusted channel confirms the request.
A general educational overview, not a substitute for professional guidance. Check current local rules and trusted primary sources before making consequential decisions.
Share this article
