You answer the phone. It’s your daughter, crying. She’s been in an accident, facing charges, and needs money now. The voice is unmistakable. Except it isn’t her.
It takes as little as three seconds of audio for AI voice-cloning tools to replicate a person with enough fidelity to fool even close family. That audio can come from a social media video, a voice note, a podcast appearance. In Florida, a mother wired $15,000 to scammers after a call from what sounded exactly like her distressed daughter.
AI scams surged 1,210% in 2025, with voice-cloning listed as a top fraud attack vector, and deepfake-enabled losses now projected to hit $40 billion globally by 2027. The corporate world isn’t immune either. A finance worker at a multinational firm was tricked into transferring the equivalent of NZ$40 million after attending a video meeting featuring what appeared to be their CFO and colleagues. It was entirely fabricated.
The most vulnerable, older adults, are disproportionately targeted, with many losing sums between $500 and $15,000 after receiving messages designed to exploit trust in family and authority figures.
This technology is being used to defeat voice-based bank security systems, generate hyper-personalised phishing emails, and build fake investment platforms complete with fabricated news articles and deepfake video endorsements. Victims are being signed-up for insurance and subscription services without their knowledge.
Businesses are also exploiting AI, not through outright fraud, but through opacity. Pricing algorithms quietly charge you more based on your browsing behaviour. Automated hiring systems screen out candidates before a human ever reads their CV. AI-powered customer service tools are designed to wear down complaints and deflect refunds.
Here in Aotearoa, the picture is particularly concerning. A Taranaki woman lost $224,000 to a scheme involving a fake video of Prime Minister Christopher Luxon, and New Zealand’s Financial Markets Authority has flagged deepfake impersonations of Winston Peters, Kiwibank CEO Steve Jurkovich, and Westpac CEO Catherine McGrath, all used to push fraudulent investment platforms carrying the logos of trusted outlets like RNZ and the NZ Herald.
The government’s response? A “light-touch, principles-based” strategy that avoids introducing any AI-specific legislation, relying instead on existing privacy and consumer protection frameworks that were never designed with deepfakes in mind. Until mid-2025, New Zealand was the only OECD country without a published AI strategy. We are playing catch-up, while the scammers run laps.
Agree on a code word with your family for emergencies. Be skeptical of any urgent call requesting money. Keep your voice off public platforms wherever you can, because in 2026, your voice is no longer just yours.
