DEEPFAKE FRAUD
Deepfake Scam Guide: How To Verify Video, Images and Calls Before You Act
Deepfakes can imitate faces, voices and public figures, but the biggest mistake is treating detection as a visual guessing game. The safer approach is to verify the request, the person and the transaction through channels the suspicious media does not control.
This guide is about verification, protection, continuity and recovery. It does not provide instructions for committing fraud or cyber intrusion.
The immediate answer
If a video, call or image asks you to send money, reveal credentials, change bank details or act urgently, stop the transaction first. Contact the supposed sender through a trusted route you already possess, verify the request independently, and do not use contact details embedded in the suspicious content.
Do not bet money on spotting visual glitches
Deepfake detection cues change quickly. Odd blinking, strange hands, lip-sync errors and unnatural lighting can be clues, but increasingly polished media may contain none of them. A real video can also look compressed or strange. Instead of asking “can I prove this is fake?”, ask “can I independently prove the request is genuine?” That shift keeps the decision safe even as generation quality improves.
Separate media authenticity from transaction authenticity
A piece of media can be genuine while the request attached to it is fraudulent, and a fake clip can contain otherwise true information. Verify the actual action being requested. If an executive appears on video asking for an urgent transfer, confirm the payment through established business controls. If a relative appears to request money, call them through a known number. Authentication should follow the money, account change or disclosure — not the emotional impact of the video.
Use a second channel that the attacker does not control
The strongest practical check is channel separation. Do not reply to the message and call the number it contains. Instead, use a saved contact, official website, banking app or another family member. For workplace requests, use the normal approval chain even if the video appears to show a senior person. If the second channel is unavailable, delay the irreversible action rather than treating the media as sufficient proof.
Be especially cautious with investment and celebrity content
Synthetic media is well suited to fake endorsements because viewers already recognise the face being copied. Treat unexpected investment promotions, giveaways, crypto pitches and miracle-product videos as advertising claims that require independent verification. Search for the offer through official sources, check regulated-firm details where relevant, and avoid sending money because a familiar public figure appears to recommend it. A convincing face is not evidence that an offer exists.
Preserve evidence without spreading it
If you believe media is fraudulent, save the URL, screenshots, account name and transaction details where safe, but avoid repeatedly forwarding the content to friends “for awareness” because that can amplify the scam. Report suspicious messages through appropriate platform and UK reporting routes. If money has been lost, contact the bank immediately. If an account has been compromised, recover the account and change credentials before focusing on proving how the media was made.
Build verification into normal life before a crisis
Families and small organisations should agree in advance how unusual requests are confirmed. That can mean a callback rule, a second approver for payment changes, a family verification phrase and a ban on changing payment details from one message alone. During a wider AI misinformation event, these boring procedures become extremely valuable. The goal is to make manipulation expensive by requiring an attacker to defeat multiple independent checks.
Continue from here
Build this topic into a wider plan, then see how the same dependency could fail in an AI-collapse scenario.
Evidence desk
These guides use current UK cyber and fraud-prevention guidance. AI-enabled voice cloning and deepfakes are treated as real fraud techniques, while any wider collapse consequences are labelled as scenario analysis.
OFFLINE FALLBACK
If identity, payments and communications fail together, paper and offline systems matter.
The 72-hour field manual covers power, communications, money, food, water and the practical information your household needs when digital systems stop cooperating.
Continue
More Cyber & Digital Resilience guides → · Communications survival →