Voice clones and fake calls: a protocol for families and finance teams.
For families and finance teams. Two short protocols you can agree in advance and rehearse, so that verifying through another channel becomes automatic.
Guide G3 · updated
How the scams work (what is documented)
- Voice clips of a relative are used in a made-up crisis to demand money. AI audio is used to impersonate someone to a bank, and real-time AI video chats pose as executives or officials 1.
- The FTC says scammers use voice cloning to make requests for money or information more believable. A familiar voice is not proof 2.
- ENISA notes that AI is notably used in vishing and online impersonation fraud 3.
- In June 2026 a Ukrainian deputy minister of digital transformation stated that social-engineering fraud is currently more dangerous than classic viruses 4.
Family protocol
- 1Agree a secret word or phrase now, and require it for any request for money 1.
- 2The rule: hang up and call back on a number you already have or look up yourself 1.
- 3Wait 10 minutes before any transfer.
- 4No gift cards or wire transfers to “family” under pressure 2.
- 5Reduce public voice and video: make social accounts private and limit the samples available 1.
- 6Teach older relatives the rule, and practise it once with a pretend call.
- 7If you have already paid, call your bank at once (use the number on your card), then report to the Cyber Police 5.
Finance-team protocol
These are NAK's recommendations, built on documented attack patterns 1.
- 1No payment, change of bank details or credential reset on the strength of a voice, video or chat message alone.
- 2Verify by callback to a number from the verified internal directory, not the number in the request.
- 3Two-person approval above a set threshold, with no “urgent” exception.
- 4Name a verifier and keep a log of every unusual request.
- 5Run short drills twice a year, including a cloned-voice scenario.
- 6If a fraud succeeds: freeze the payment route, ask the bank to recall the transfer, preserve evidence (see the guide “The first 24 hours after a deepfake”), report to the Cyber Police 5, and inform CERT-UA if systems were accessed 6.
Ready-made wording
- Callback: “Thank you. I will call you back in a few minutes on the number in our directory.”
- Safe word: “What is our safe word?”
- Message to staff after an incident: “Today someone tried to defraud us in the name of [name]. No payments or changes of bank details without a callback and a second signature. Report any similar request to [named person] immediately.”
Limits
This is a guide, not legal advice and not a service. NAK cannot verify your file, recover money, remove content or investigate. This text was last checked on 22 Oct 2026; the sources used are listed below.
Sources
- 1FBI Internet Crime Complaint Center. Criminals use generative AI to facilitate financial fraud (PSA I-120324) · accessed 2026-10-22
- 2Federal Trade Commission. Family emergency scams · accessed 2026-10-22
- 3European Union Agency for Cybersecurity (ENISA). ENISA Threat Landscape 2025 · accessed 2026-10-22
- 4Cabinet of Ministers of Ukraine. Mintsyfra: how to protect bank cards from fraudsters and recognise a deepfake (29 June 2026) · accessed 2026-10-22
- 5Cyber Police Department, National Police of Ukraine. Cyber Police: incident form and appeals · accessed 2026-10-22
- 6Cabinet of Ministers of Ukraine. How CERT-UA responds to cyber incidents, from report to recovery · accessed 2026-10-22
Read next.
- Guide G2How to check a suspicious video, voice message or call
- Guide G1The first 24 hours after a deepfake of you or your company
- Guide G6Protecting your accounts and digital identity
- Guide G7Cyber preparedness minimum for NGOs and SMEs
- Focus areaDeepfakes & Synthetic Media
- Focus areaCyber Warfare Resilience