Imagine opening Instagram and seeing a video of hundreds of people fighting to enter a supermarket. The caption says: "People are panic-buying because the country is running out of food." The video looks real. The people are real. The supermarket is real. So is the story true? Not necessarily. In Open-Source Intelligence (OSINT), one of the most important skills is separating what a video actually shows from what someone claims it shows.
Sometimes the truth needs nothing more than a screenshot, a search engine, and the patience to question what everyone else has already accepted.
The Aldi video that fooled people
During the early COVID-19 pandemic, a video circulated claiming to show panic-buying outside an Aldi supermarket in Haarlem, the Netherlands. Crowds were gathered outside a store — and people shared the clip as evidence. Investigators at Bellingcat took a different approach: instead of trusting the caption, they searched for the footage itself. A screenshot from the video went into reverse image search — and surfaced a German video titled "ALDI Sonderverkauf in Kiel": a special sale in Kiel, Germany, uploaded to YouTube in 2011, years before COVID-19 existed.
The original footage was real. The interpretation attached to it was not. No fabricated images, no AI-generated people, no obvious editing — yet still misinformation, simply because someone changed the context. We keep asking whether a video is fake. The more useful question is whether the story attached to it is true.
Step 1: Find the earliest version of the video
First question, every time: where did this video appear before it went viral? An account may have uploaded the clip yesterday — that doesn't mean the event happened yesterday.
- Capture several screenshots from different moments — frames showing distinctive buildings, signs, vehicles, or clothing.
- Search them with Google Lens, Bing Visual Search, and TinEye. Don't rely on one screenshot: cropping, mirroring, subtitles, and overlays make the same footage harder to match.
- Search distinctive caption phrases with "original video," "old footage," or the claimed location's name.
Step 2: Investigate the location
A viral video allegedly shows an explosion in a particular city? Look at the background before accepting it. Street signs. Building shapes. Mountains, road markings, shop names, landmarks — compared against satellite imagery in Google Maps and Google Earth. This is geolocation: determining where footage was actually recorded. Bellingcat's conflict-footage investigations show how road markings and small environmental details can establish where a video was filmed — and expose misleading location claims. But finding the right city is only half the job: a video recorded in the right place can still be years old.
Step 3: Check when it was recorded
Footage genuinely showing a military vehicle in Ukraine doesn't prove it was filmed during the latest fighting. In a documented case, military training footage from Russia's 2012 Kavkaz exercises circulated as if it showed fighting in eastern Ukraine in 2016 — real soldiers, real weapons, real explosions, completely unrelated to the event in the caption.
To investigate timing: compare the earliest uploads you can find, visible weather, shadows, seasonal vegetation, and any dated information in-frame. And remember — a platform's upload date tells you when a file was posted, not when it was recorded. Metadata can be missing or altered.
Step 4: Look for independent confirmation
One account posting a video is not confirmation. Neither is twenty accounts reposting the same original clip. Look for credible news reporting, statements from relevant authorities, eyewitness accounts, and other footage that independently supports the claim — and check whether sources offer genuinely separate evidence or simply repeat the same unverified post. Major explosion? Ask whether other recordings exist from different angles, whether the location is confirmed, whether reliable reporting supports the date and event.
Sometimes the evidence isn't enough for a firm conclusion. That's not a failure — it's a reason to keep the claim unverified.
Step 5: Don't let AI detection tools do your thinking
AI-generated video is getting convincing, and it's tempting to upload a suspicious clip to a detector and accept whatever comes back. Be careful: detectors produce false positives and false negatives, and compression, editing, low resolution, and reposting all muddy the analysis. Instead, investigate the evidence around the video — inconsistent reflections, unnatural movement, mismatched audio, suspicious edits, contradictions between caption and independently verified facts. None of these proves manipulation alone. Even a video with zero visible AI artifacts can be misleading, because it was recorded somewhere else or at another time.
Slow down before you share
OSINT isn't about proving everything online is fake. It's the discipline of separating evidence from assumptions. The Aldi case shows how an ordinary screenshot can uncover a misleading narrative; the military-footage case shows how old material gets repackaged as breaking news. Both reveal the same weakness: people trust the explanation attached to a video before investigating the video itself. My personal rule: never confuse a convincing image with a verified claim.
Three questions before sharing a viral video
- Where did this footage originally appear?
- Can I independently verify its location and date?
- Does reliable evidence support the story attached to it?
You don't need expensive software or advanced technical skills to begin — just curiosity, patience, and a willingness to change your mind when the evidence points somewhere unexpected. Because the most valuable OSINT discovery isn't always finding a fake video. Sometimes it's discovering that a real video has been telling the wrong story all along.
🔄 The flip side of verification: responsibility
These skills protect you as a viewer. They also bind you as a creator. AI video tools now put production power in everyone's hands — including e-commerce sellers using purpose-built platforms like VEONIB (veonib.com) to turn real product links into marketing videos. The OSINT lesson applies in reverse: keep the story your video tells grounded in what the product actually is. In a feed where audiences have learned to verify, honest context isn't just ethics — it's what makes marketing survive contact with a skeptical audience.
Let the video tell the product's true story
One product link. The AI understands the actual product,
and builds the video from what's real.
Visit veonib.com and turn your first product link into a video