How to Check Whether a Photo or Video Is Real
Edited and generated images have always existed, but tools that create convincing fakes in minutes have made casual scepticism more necessary. Perfect certainty is rarely possible for an ordinary viewer. What you can do is combine several quick checks, weigh the context and avoid sharing material you cannot back up.
Begin with the source, not the pixels
Ask who posted it and whether they were in a position to capture it. A first-hand account from an identifiable person or established news organisation carries more weight than an anonymous repost. Search for the same event reported elsewhere: genuine incidents usually appear in several independent places, with different angles.
Reverse image search
Search engines and dedicated services let you upload a picture, or paste its address, to find earlier appearances. Often an old photo is being passed off as a new event. If the earliest version has a different caption, date or location, the claim is likely false. For video, take screenshots of a few distinct frames and search with those.
Look for visual inconsistencies
- Hands, teeth, ears, jewellery or glasses that look warped or merge into the background.
- Text on signs or clothing that is garbled or changes between frames.
- Shadows and reflections that disagree with the apparent light source.
- Repeating textures or oddly smooth skin.
- In video, lip movement that drifts from the audio or an unnatural blink pattern.
Be cautious with this list. Generation methods improve quickly, and genuine photos can also contain compression artefacts and strange details. A flaw is a reason to investigate, not a verdict, and the absence of flaws proves nothing.
Check the surrounding facts
- Does the weather, clothing, vehicle style or signage fit the claimed place and time?
- Do landmarks match satellite or street-level maps?
- Is the story consistent with what official sources or local reporters say?
- Does the content seem designed to provoke strong emotion or immediate sharing?
Metadata and provenance
Image files can contain embedded details about the camera, time and sometimes location, but social platforms often strip this on upload, and it can be edited. Some organisations are adopting content credentials, a label attached to a file that records how it was made and edited. Where present, they are helpful, but absence does not mean something is fake, since the system is not yet universal.
Detection tools have limits
Online services claim to identify generated images and videos. Their answers can be wrong in both directions, and a confident score can mislead. Use them as one input among several, never alone.
Fact-checking resources
Many countries have independent fact-checking groups and newsroom verification teams that publish explanations of viral claims. Searching a claim plus the words “fact check” is a quick way to see whether others have already examined it.
A short routine
- Pause before reacting or sharing.
- Identify the original source.
- Run a reverse search.
- Look for corroboration.
- If still uncertain, say so, or do not share.
Admitting uncertainty is a perfectly respectable position. Not spreading an unverified clip is the most helpful action many of us can take.
Be kind to those who get fooled
Anyone can be taken in by a convincing fake, including careful people. If a friend shares something dubious, reply with the checks you ran rather than mockery. Gentle corrections keep conversations open and make people more willing to pause before passing material on next time.