Safety and compliance
Deepfakes and voice fraud: protecting your business
A call that sounds like your boss, a video call that looks real. How to protect your business against fraud using imitated voices and images.
5 min read
With AI, someone can convincingly imitate a voice or face from short recordings, for example a video on social media. Fraudsters use this to get employees to make payments or share data, with a call that sounds like the owner or a video call that looks genuine.
What it looks like
- An urgent call from "the boss" asking for a payment from abroad.
- A voice message from a "colleague" who needs a code.
- A video call with a "supplier" passing on a new account number.
Why it works
It plays on trust and time pressure. The voice is familiar, the story is plausible and there's a rush.
Rules that protect you
- Payments above a set amount always require a second person and a check via a fixed channel.
- A code word for urgent requests by phone, known only internally.
- Call back on a number you already have, never the number that called.
- No exceptions under time pressure. The real owner will understand that you check.
- Report it: every suspicious attempt gets reported, even if nothing happened.
What you share online
Videos and audio of managers on social media make imitation easier. You don't have to stop communicating, but be aware of the risk and make sure the rules above are known.
Checklist
Your answers stay only in this browser.
Prompt to copy
We're a small business with [number] staff. People authorised to make payments: [x]. Write a short procedure of no more than one page to protect us against fraud using imitated voices or video: when an extra check is needed, how we check and how we report. In plain language for all staff.
