Real or not? Top 5 ways to spot ‘deepfake’ videos and images
More and more AI deepfakes are cropping up. From Taylor Swift to Pope Francis, no one is immune to the AI fakery which is quickly becoming one of the biggest problems confronting us online. Video and image generators like DALL-E, Midjourney and OpenAI’s Sora are making it easy for people without any technical skills to create deepfakes which might seem harmless but can actually lead to scams, identity theft, propaganda and even election manipulation.
Deepfakes usually have obvious errors, like hands with six fingers or eyeglasses that have differently shaped lenses but since AI has improved a lot, you can also look out for unnatural blinking patterns among people in deepfake videos, an electronic sheen to the skin and also check consistency of shadows and lighting.
One of the most common deepfake methods is face-swapping which is why you need to look closely at the edges of the face in order to identify deepfakes. In a video, look at the person’s mouth while speaking and check if their lip movements match the audio perfectly. You can also look at the teeth and see if they are blurry or consistent with how they look in real life.
When spotting a deepfake video, always check what is the context of the video or the photo. Simply put, if you see a public figure doing something that seems exaggerated, unrealistic or not in character” it could be a deepfake.
Yes, use AI to fight AI. There are various tools that can be used to analyze photos or videos like Microsoft has developed an authenticator tool that can help. Intel’s FakeCatcher can also be used as it has algorithms to analyze an image’s pixels to determine if it’s real or fake.
