How to check photos and videos in the news

By Omid Jamshidi · Published 26 October 2026 · 5 min read

Old pictures in a new context, edited videos and AI-generated images – simple steps anyone can take to check where a photo or video comes from before believing or sharing it.

Pictures are powerful: we tend to believe what we see. That is exactly why photos and videos are among the most frequent forms of misinformation – especially during wars, disasters and protests, when emotions run high and information is scarce. Most misleading images are not sophisticated fakes but real pictures shown in the wrong context. This guide explains simple checks that anyone can carry out in a few minutes.

The most common problems

• Old images in a new context: a photo from an earlier war, another country or a different year is presented as current. This is by far the most frequent case.
• Misleading captions: the picture is real and recent, but the text claims something it does not show – a different place, different people, a different cause.
• Edited images and cut videos: parts are removed, added or rearranged; a video clip leaves out what happened just before or after.
• Generated images: pictures or videos created entirely with AI tools. They are becoming more realistic and harder to recognise.

Step 1: Who published it first?

Before looking at the image itself, look at its source. Reputable media name the photographer or agency (for example Reuters, AFP, AP, dpa or EFE) and explain where and when a picture was taken. An image that circulates without any source on social networks – "shared 50,000 times" – has not been checked by anyone just because many people forwarded it.

Ask: Is the account that posted it known and accountable? Did it take the picture itself or only pass it on? Do other, independent sources show the same scene?

Step 2: Reverse image search

A reverse image search shows where else a picture has appeared on the internet – and often when it appeared first. If a "photo of today's protest" already existed five years ago, the case is solved.

How it works:

1. Save the image or copy its address. With a video, take a screenshot of a characteristic moment.
2. Upload it to a reverse search service such as Google Lens, Bing Visual Search, TinEye or Yandex Images. Different services find different results, so trying more than one is worthwhile.
3. Look for the oldest appearances and for versions with a larger size or a caption from a reliable source.

For videos, journalists also use free tools such as the InVID-WeVerify browser plugin, which splits a video into key frames that can then be searched one by one.

Step 3: Check the details

Many pictures contain clues about where and when they were taken:

• Signs, shop names, licence plates and language – do they match the country claimed?
• Landmarks, buildings and landscape – can they be found on a map or in satellite images?
• Weather, vegetation and clothing – do they match the season and place? Snow in a city where it is summer is a warning sign.
• Shadows and light – do they fit the time of day that is claimed?

Hidden file information (metadata) can also contain a date and a location, but social networks usually remove it, and it can be changed. It is a hint, never proof.

Step 4: Generated and manipulated images

Earlier, AI-generated images often betrayed themselves through deformed hands, nonsense writing or strange details in the background. Such errors are becoming rarer, so the absence of visible mistakes proves nothing. More reliable questions are the same as above: Who published it? Does it appear in any trustworthy source? Is the scene confirmed by other pictures from different angles?

Some cameras, image programs and platforms now attach Content Credentials – a kind of digital label that records where an image came from and how it was edited. Where they exist, they are useful; their absence, however, does not mean that a picture is fake.

Step 5: Use fact-checkers

Many organisations check viral images and claims professionally and publish their results, for example AFP's fact-checking service, dpa-Faktencheck and Correctiv in Germany, Maldita, Newtral and EFE Verifica in Spain, Full Fact in the UK, PolitiFact in the US and Factnameh for Persian-language content. Searching for a few key words of the claim together with "fact check" often leads to an existing check within seconds.

When in doubt: wait

In breaking situations, the first pictures are often wrong or taken out of context, and corrections come later. A few hours of patience prevent most mistakes. If you cannot verify an image, do not share it – or say clearly that its origin is unknown. Every share makes a false picture harder to stop.

Pictures and Omid57.info

Omid57.info shows headlines only, without photos or videos, and links every item to the original article. Pictures belong to the outlets that publish them and are their responsibility. When the same event appears in several outlets on our site, opening two or three of them is an easy way to compare whether their pictures match. Our guide on how to read the news critically has more checks for the text of a story, and how news agencies work explains why so many outlets use the same agency photos.

Conclusion

Most false pictures are not clever fakes but old or real images with a new, wrong story. Asking who published a picture first, running a reverse image search and looking at a few details uncovers many of them within minutes. And when a check is not possible, the safest choice is simple: do not share it yet.