How Do Scammers Use Old Interview Footage to Build a Deepfake?
Many crypto deepfakes start with real, public interview footage rather than fully synthetic video — making them harder to dismiss. Here's how the technique works.
Rather than generating an entirely synthetic video from scratch, many crypto deepfake scams start with genuine, publicly available footage of a real person and manipulate it — a technique that's both more efficient to produce and considerably harder to dismiss as obviously fake.
Why Starting With Real Footage Is More Effective Than Generating From Scratch
Deepfake technology generally works by training on and manipulating existing video and audio of a target — the same underlying training data requirement covered when discussing voice cloning applies here — meaning genuine, high-quality source footage produces more convincing results than attempting to generate an entirely synthetic person.
Where This Source Footage Typically Comes From
Public interviews, conference talks, podcast appearances, and other widely available video of well-known crypto and business figures provide ample raw material — the more publicly available footage of someone exists, the easier it becomes to produce a convincing manipulation using their likeness.
How the Manipulation Technique Typically Works
Rather than fabricating an entirely new scene, scammers can take a genuine video clip and alter specific elements — swapping in fabricated audio synced to match the original mouth movements, or modifying the video to change what the person appears to be saying, while keeping the surrounding footage, setting, and overall context genuinely authentic.
Why This Hybrid Approach Is Particularly Deceptive
A video where the setting, background, and general appearance are all genuinely real — because they're drawn from actual footage — while only the specific spoken content has been altered, is considerably harder to dismiss than an obviously synthetic scene, since so much of what's visible checks out as authentic.
Why Out-of-Context Editing Doesn't Even Require Full Deepfake Technology
A simpler, related technique involves selectively editing or clipping genuine footage to make a real statement appear to mean something entirely different than its original context — not technically a deepfake at all, but achieving a similar deceptive effect through editing alone rather than synthetic generation.
Why Checking the Original Source Interview Can Reveal the Manipulation
Searching for the original, complete interview or footage that a suspicious clip appears to be drawn from can reveal whether the specific statements attributed to the person in the circulating clip actually match what they said in the genuine, complete original context.
Why Reverse Video Search Tools Can Help Identify the Source
Some tools allow searching for where else a specific video clip or its component footage has appeared online, which can help identify the original, unaltered source and reveal discrepancies between the original and the circulating manipulated version.
What to Check When Encountering a Suspicious Clip of a Known Figure
Searching for the original, complete source footage the clip appears to be drawn from, and comparing what the person actually said in that original context against what the circulating clip claims or implies.
Search for the original source footage before trusting a suspicious clip's claimed statement — genuine footage manipulated with fabricated content is often more convincing than fully synthetic video.