A video of a celebrity recommending an investment, a video call from 'the CEO' asking for an urgent transfer, a voice message from a relative in trouble. Any of these can now be entirely fake, the product of AI. In May, for example, Israel's National Cyber Directorate warned of a scam using deepfake videos of Chief Rabbi David Yosef to lead people to sites that steal payment details. So how do you spot them?
Visual signs that sometimes still help: lip movements that do not quite match speech, unnatural blinking or no blinking, blurry edges around hair, ears or glasses, lighting on the face that does not match the background, and teeth or hands that look odd. In audio, listen for overly flat intonation, missing natural breaths, or pauses in strange places. On a video call, ask the person to turn their head sideways or pass a hand in front of their face; some tools still struggle with that.
But let's be honest: the technology is improving fast and these signs are disappearing. Relying on your eyes alone will eventually fail. So the best defense is not spotting the fake but spotting the scam. Scams almost always share the same ingredients: urgency ('now'), secrecy ('don't tell anyone'), money or personal details, and contact through an unusual channel.
Questions to ask every time: would this person contact me like this, through this channel? Is there a real reason for urgency? Can I check the request another way, for example by calling a number I know? If a celebrity recommends an investment in an ad, did they also say it on their official channels? In most cases, thirty seconds of checking is enough to expose the scam.
And if you come across a deepfake? Do not share it, even to warn others, since sharing only spreads it. Report it to the platform where you saw it, and if it is a scam, report it to the National Cyber Directorate's 119 hotline. And to protect your family, set a family code word: a question only you know the answer to. No AI knows a secret that was never said out loud.
