You just received a photo on WhatsApp, Facebook or Twitter. The image makes you angry, sad or joyful, and the caption encourages you to share it as widely as possible. You're a little cautious, however, because the story seems too good to be true. You are right to be careful. Here are a few tips for verifying images and tracing a photo’s origin on your own.
There's nothing like a compelling photo to capture someone’s attention on social media. Full of emotion, a photo can compel a reader to linger on an article, click on a link and share content widely. Journalists know this and spend time choosing just the right photo to illustrate their articles.
Unfortunately, people who spread "fake news" also understand the power of a photo. To generate a maximum number of clicks (and therefore money), some people manipulate or misappropriate photos that have nothing to do with the topic in question... just to get your attention.
Identify old images that are being used out of context
There are a lot of digitally altered images circulating online. That said, it still requires some technical skill to alter an image. However, there is a much more simple way to trick people using photos: you can find an old image and alter the caption, making the photo tell whatever story you want.
The best reflex to have when you see an image you think was taken out of context? Doing a reverse image search.
Let’s look at an example. This photo shows a woman threatening a police officer who is pointing a pistol at a man on the ground. The caption claims that this photo shows a mother in the Dominican Republic trying to protect her son from a police officer. The story is touching, the photo is high-quality and it is a great shot. In fact, it is perhaps too great of a shot – and that makes us suspicious.
Here’s how to verify a photo with a reverse image search.
1) Start by copying the address of the photo by right-clicking on it (or by pressing it for a long time on your smartphone.)
2) Then go to Google Images and paste in the address.
3) Click on “search by image” and check out the results
The first webpage in the Google results comes from 2015. If you click on it, you can read that the photo is actually an image from a movie called "Cristo Rey". To confirm that, we can search for the movie on the film website IMDb. Sure enough, the image appears in the gallery of the movie's page.
Find out more about this hoax by reading our article on the topic.
… But be careful, because even Google makes mistakes
Google Images was the first online tool that people used widely to carry out reverse image searches. However, it isn’t perfect and it doesn’t always find the origin of an image. Sometimes, it can even provide false information.
Take the example of the photo below. FRANCE 24 blurred this image because it shows charred bodies being examined by workers from the Red Cross. The photo is often misappropriated by social media users who claim, for instance, that it shows Christians massacred by Boko Haram in Nigeria.
If you do a reverse image search on Google, this is what it pulls up:
According to Google, the photo was taken during a massacre in Duékoué, Ivory Coast, which occurred during the crisis that swept the country between 2010-2011.
However, if you continue searching, you’ll see that, in reality, this photo has nothing to do with the massacre in Duékoué. It was actually taken in July 2010 in the Democratic Republic of Congo, when a fuel truck exploded.
So why does Google get it wrong? Simply because the Google algorithm looks for the most probable context for an image by seeking out articles containing the photo.
However, this photo has been widely misappropriated; it has been used by too many articles or websites that claimed it was a photo of Duékoué. Because of the high frequency of these articles, it has morphed into the truth, at least for Google.
Other tools for doing reverse image searches
If Google doesn’t give you a satisfactory response, then it’s time to try your luck with other tools so you can do a double or even triple verification of a suspicious image.
Check out a few other sites below:
Yandex, a Russian search engine
If you're looking to find the source of an image that seems to come from Russia or Eastern Europe, Yandex is your best bet. Its algorithm brings up different webpages than Google.
Its reverse image search tool is also particularly good at identifying faces.
TinEye, an independent verification tool
TinEye is one of the oldest reverse image search tools. It allows you to use filters to, for example, search for the oldest photo, the largest file size or the most modified image.
While it might not be the most efficient search option for recent images that pop up in the news, it does have functions that can help you spot some photoshopped manipulations in photos.
Baidu, the Chinese search engine for images
Baidu is a great place to start if you're looking to match up a photo or video with something on the Chinese web.
Most of the results from other search engines don't show content from Chinese sites. But by focusing your search with Baidu, you can sometimes find articles or photos similar to the ones you're investigating.
Here's an example of how we used Baidu to analyse an image of children dressed up in costumes during the Covid-19 pandemic in China.
Plugins can help you carry out a search in one click
There are two different plugins you can install directly into your browser to allow you to search for an image easily.
You can download RevEye for Google Chrome or Firefox in order to simultaneously search with the four reverse image tools we mentioned above.
You can also use the InVid WeVerify plugin to do the same thing, simply by right-clicking on the image.
Finally, keep in mind that no tool is perfect and none used alone will allow you to definitively identify the origin of an image. If you have doubts about a particular image, sometimes it’s better just not to share. You don't want to get it wrong and help to spread a hoax.
And if you can’t verify an image yourself, then don’t hesitate to contact us via Facebook or email and we’ll have a look!
