Protecting yourself from cloned-voice scam calls: effective strategies
Voice cloning has made fraud more sophisticated. A few clean seconds can reproduce speech convincingly, threatening individuals and businesses. We explain how to recognise fake calls, why families are targeted and how to reduce risk.
Voice cloning and threat scale
Current software can analyse just 3-4 seconds of audio and reproduce intonation and timbre with approximately 85% accuracy. Training on several recordings can raise that to 95%. Public videos, voice messages and recorded calls can provide source material, making careless sharing a security risk.
A major telecom operator reports cloned voices in around 10% of fraudulent calls. Investigations have identified thousands of victims of audio and video deepfakes.
Why voice fraud targets families
Fraudsters exploit emotion and urgency. Hearing a relative ask for help can prompt action without checks. Surveys suggest around half of people would respond to a supposed family member's request, especially in an emergency.
Fraudsters simulate emergencies and discourage time for verification.
Recognising a cloned voice
- Unnatural speech: unusually flat speech and long pauses;
- Repeated phrases: evasive answers and repetitive wording;
- Video-call synchronisation problems: unnatural blinking or speech synchronisation;
- Pressure and urgency: pressure not to hang up or verify;
- Departures from familiar behaviour: refusing to pause or verify independently.
Protecting people and businesses
Combine technical and organisational safeguards:
- Limit public voice recordings: limit public clean-voice recordings to reduce cloning opportunities.
- A family or business verification code word: agree a private verification phrase not shared publicly.
- Call back on a known number: call back independently to verify urgent requests.
- Train employees and family members: explain fraud warning signs and verification procedures.
- Systematic marketing and analytics: transparent analytics can reveal communication and SIP-traffic anomalies that signal fraud.
Marketing and security work together
AI expands fraudsters' capabilities but also provides tools for detecting and preventing threats. Connecting strategy, analytics and AI supports both growth and security. Integrated digital marketing must account for these risks to reduce losses and build resilience.
