TL;DR
Someone in Ards and North Down has lost £250,000 to a fraud that began with a fabricated video of a recognisable figure from the financial world promoting an investment. Northern Ireland’s police service has described the operation as sophisticated, and it did not stop at the video.
How the fraud actually worked
The synthetic endorsement was the entry point, not the mechanism. It advertised an investment promising substantial returns, and the victim started with a modest sum — small enough to feel like a test.
What followed was a long game. Contact moved to WhatsApp. The fraudsters spent time building confidence before pushing the amounts upward, had the victim open multiple online accounts, and took remote control of their computer on the pretext of assisting with the investment. The victim was eventually encouraged to borrow in order to keep investing. Money went to accounts the criminals controlled.
That sequence matters more than the deepfake itself. The video bought credibility at the start; the losses were built by ordinary social engineering, applied patiently.
Why the endorsement worked
Police did not name the personality whose likeness was used. BBC News NI has separately seen scams online deploying figures including Steven Bartlett, Peter Jones and Martin Lewis — people whose public reputation is built specifically on being trustworthy about money.
The PSNI’s guidance is direct: an apparent celebrity endorsement is not evidence that an investment is real. The force notes that fraudsters can now produce highly convincing adverts, websites and communications, with AI generating the videos, the voices and the endorsements themselves.
Looking forward
Superintendent Joanne Gibson, who chairs the Scamwise Partnership, asked people to “stop and check” before acting on financial approaches, and to raise the subject with relatives. The formal advice is to verify independently who you are dealing with, and to check any firm’s FCA authorisation before parting with money.
For UK businesses, the case is a marker of where synthetic media has landed. Detection guidance still tends to assume a viewer might spot the fake. This victim had no reason to look — the video was doing the job of an introduction, and everything that extracted the money came afterwards, through channels no deepfake detector inspects.