AI voice cloning used in WhatsApp impersonation scams
Scammers are using AI voice synthesis tools to impersonate family members in WhatsApp fraud schemes, reportedly stealing approximately half a million pounds from UK victims since early 2025. Fraudsters send messages from unknown numbers followed by synthetic voice messages that mimic real family members' voices to make requests for urgent money transfers.
What happened
Since early 2025, scammers have been using artificial intelligence voice cloning tools to impersonate family members in WhatsApp fraud schemes. Victims receive messages from unknown numbers claiming to be children or close relatives with phrases like 'Hello mom' or 'Hello dad', followed by stories about lost phones or urgent money needs.
To increase credibility, fraudsters send synthetic voice messages created using AI tools trained on videos found online, generating audio that closely mimics the voices of victims' family members. The scammers request rapid money transfers to new accounts, claiming urgent bills or emergencies.
Reported impact
Approximately half a million British pounds have reportedly been stolen through these scams in the UK since the start of 2025. Global cybersecurity experts noted that voice forgery is now possible using commercially available tools, with one security consultant reporting he was able to deceive his own mother with a synthetic voice created in minutes.
Index analysis
The scams exploit the trust victims place in their family members' voices and the difficulty of distinguishing synthetic from genuine audio. WhatsApp uses end-to-end encryption to protect conversations but displays warning notifications for messages from unknown numbers; however, users may still be vulnerable to social engineering tactics.
Governance controls
WhatsApp advises users to verify phone numbers through alternative channels when receiving unexpected contact, use video calls to confirm requests before transferring money, and enable two-factor verification. Security experts recommend caution with unexpected voice messages and avoiding clicking links from unregistered contacts.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.