The rapid development of AI offers us many benefits and conveniences, but it also presents a host of new challenges. It’s becoming increasingly difficult to determine whether the photos and videos you see online are real or have been created using AI. In some cases, you can probably tell right away that it’s an AI video—for example, if you see Mark Rutte singing opera—but AI can also be used to manipulate you with photos or videos that look just like the real thing. In this blog, we’ll give you 5 tips for spotting deepfakes!
What are deepfakes?
Deepfakes are fake photos or videos that look remarkably real. Often, these photos and videos are virtually indistinguishable from the real thing, which means they can cause a great deal of harm. To create deepfakes, people use deep learning, a form of machine learning. They duplicate someone’s voice or facial expression and superimpose it onto another recording or photo. This makes it seem as though someone was present at a location where they could never have been. You can imagine that this can lead to a lot of deceptive practices, since many people cannot easily recognize deepfakes.
How do people create deepfakes?
There are now plenty of apps and other types of software available that let you create deepfakes, even without having to pay for them. In addition, there are websites that offer deepfake software, making it easy to create these kinds of videos. Currently, artificial intelligence is primarily used to create fake videos or photos. This technology is based on machine learning, and the computer model is constantly fed with data. As a result, it can sometimes be very difficult to spot deepfakes.
How can you spot deepfakes?
There are several ways to spot deepfakes and protect yourself from misinformation and manipulation:
- Unknown sender: If the video comes from a reputable organization, such as a newspaper or a local broadcaster, there’s a good chance the content is genuine and not a deepfake. If the video or photo comes from an unknown sender, don’t just assume that what you’re seeing is real. Do some research on the source first.
- Check out different sources on a topic: Is a photo or video about a topic that’s currently trending and about which people have strong opinions? Don’t base your opinion on just one source; instead, look for more information from various perspectives. This way, you’ll have a complete picture of the situation before forming an opinion. Can’t find any further information about the images in a video or what you see in a photo? Then there’s a good chance it’s a deepfake!
- It’s all in the details: It’s often extremely difficult to spot deepfakes because they differ very little from real footage. Still, there are a number of details that can give a deepfake away. For example, do facial expressions look unnatural or do they not match what is being said? Do people have extra fingers, or are fingers missing? Does the person in the video never blink? These are all details that can indicate that a video or photo is a deepfake.
- Using Google Images to verify: Want to check whether a photo or video is real? You can use Google Images for this. Take a screenshot of the image or video and run it through Google Images. This will help you identify the original source of the content, allowing you to verify whether the video or photo is the original. Google has also recently added features to its search engine to combat AI-generated images.
- Pixels and waves: Check for stuttering pixels or blurry waves in a video. If you see these, it’s likely a deepfake.
Conclusion
Spotting deepfakes isn’t always easy, but if you know what to look for, you can effectively protect yourself against misinformation. Pay attention to the source of the video or photo, to details that don’t add up or seem a bit off, to pixelation and blurring, and to facial expressions that don’t match. Finally, be sure to do your own research if something seems suspicious!