- Alex Pack
How to spot fake images and videos online
A brief explainer on the "digital breadcrumbs" attached to images and videos, that you can use to verify their stated background
In the immediate aftermath of an attack, everyone on your #Facebook, #Twitter, and #LinkedIn becomes an expert and begins sharing photos and videos of the event, perpetrator, or victim that they found through "research" online. But, how do you know that those photos are genuinely related? Because your uncle shared a post? You got sent the file from a friend in a WhatsApp or GroupMe?
Photoshop, the well-known Adobe program that allows people to edit photos, was first released more than 30 years ago. In intervening time, the technology for creating partially or wholly fraudulent images and videos has only continued to improve. Now, even those with only rudimentary skills in editing software can make believable fake photos or videos.
A totally* real photo of me with the Eiffel Tower
Dangers of fakes
While some doctored photo and videos are innocent (like those fun apps that let you make Eisenhower sing soul--wouldn't that be a fun video to watch?), there are serious dangers to fake images and videos. The main concern with any of these fraudulent pieces is disinformation.
While we know that disinformation is prominent across the social media environment, studies have found that videos and photos are more compelling and believable, especially when related to the news or a breaking development. One study published in the Journal of Computer-Mediated Communication found that a majority (58%) of participants in their study were likely to believe a fake video they saw, compared to less than half (48%) who heard the information, and approximately one-third (33%) who read the information.
And this spread of digital disinformation has very real world consequences. For example, the mob riots and murders in India in 2017-2018 were partially sparked by fake videos about child kidnapping.
One potential avenue to assess the veracity of images and videos are "digital breadcrumbs" attached to each file--metadata. This data is included in any image or video file regardless of the digital device used. These "digital breadcrumbs" are created at the same time as the file, and remain accessible (at least until it is removed by someone or something).
Generally, metadata includes the image size, device used to capture the image, date and time the image was captured, and a general location that the photo was taken (typically expressed as longitude and latitude coordinates). Below, I used a well-known online tool to pull up the metadata on a photo that I took while traveling in Vietnam.
Image taken while traveling in Vietnam
Metadata for the image
The online tool was able to identify that I took the photo on my iPhone 6, on 29 December 2019, and was able to provide the longitude, latitude, elevation, and even the direction that the camera was facing when I took the photo. When checking the coordinates against Google Maps, the data was accurately placed around 100 meters.
Given this accuracy of identifying locations, times, and tools used, it is clear that metadata is a great source of information for determining the validity of an image in context. The question is how? The answer comes back to the "digital breadcrumbs" analogy. Like the breadcrumbs used by Hansel and Gretel to find their way home, "digital breadcrumbs" in the form of metadata can be used to follow images and videos to their original source. Ultimately, by being able to access the metadata, you can verify if the claimed date/time/location of a file is accurate.
Limits of metadata?
While metadata is a powerful tool to assess the legitimacy of a files claimed provenance, it is not without limits. It is relatively easy to remove or manipulate metadata with readily available online and desktop tools. For example, one of the most widely available metadata reading tools offers the ability to immediately edit or modify an images metadata. By using this tool, it is easy to adjust the file to make it appear that it was taken by a different device, in a different location, and at a different time.
Another significant limitation of metadata is its availability in the social media ecosystem. Unfortunately, when images or videos are uploaded to certain social media sites (such as Facebook and Twitter), the platform automatically strips much of the embedded metadata. As a result, attempts to capture metadata from files on these platforms may be more difficult. While there are some platforms that maintain the metadata, it is generally harder to pick up the "digital breadcrumbs" on the major social networks.
Ultimately, while metadata is a major asset for assessing the veracity of an images claimed background, it is not enough by itself. Instead metadata should just be one of multiple tools for determining if an image or video should be trusted in our increasingly digital world.
Be careful what you share!
It is unbelievably tempting to share photos that you see online, especially when the topic is emotionally charged or #trending.
When you see videos or photos you often feel you are seeing objective truth play out--"seeing is believing," right? And you, being the good person you are, can't just sit by as you watch horrors or injustice play out. So, your first instinct is to click re-tweet or share, letting all of your followers in on what you just saw.
Unfortunately in today's world, more often than we care to admit, photos can be edited and video can be taken out of their original context or wholly falsified. This is especially true in the aftermath of an attack. So as you watch videos or photos streaming onto your social media feeds, be careful what you choose to share and amplify, because while "a picture is worth a thousand words," spreading a fake one can spread a lie to thousands of people.