5 Ways to Tell If a Video or Image Is Fake, Open your browser, navigate to a suspicious image or video, and apply these simple checks. The rise of deepfake technology has made it harder than ever to trust what we see online. Scammers now use incredibly realistic fake videos and images to trick people, especially in investment and romance scams.
These aren’t just blurry fakes anymore. Modern deepfakes can convincingly show someone saying or doing things they never did. Knowing how to spot them is a crucial skill in protecting yourself and your money.
What is a Deepfake?
A deepfake is a type of media-usually a video or image-that has been altered using powerful artificial intelligence (AI). The “deep” comes from “deep learning,” a type of AI. This technology can swap faces, make people say things they didn’t, or even create entirely new scenes that look real.
In the UK, these fakes are increasingly used in fraud. Imagine a scammer creating a fake video of a famous investor endorsing a dodgy scheme, or a fake video call from a supposed romantic partner asking for money. The emotional impact is huge, making people more likely to fall for the trick.
Check for Inconsistencies (Eyes, Skin, Lighting)
One of the easiest ways to spot a deepfake is to look for things that just don’t add up. Pay close attention to the person’s face. Are their eyes blinking naturally? Deepfake eyes sometimes blink too little, too much, or in an unnatural way.
Also, check their skin. Does it look too smooth, or perhaps too rough in certain areas? Sometimes the skin tone might not match the rest of their body, or it might have an odd texture.
Lighting is another big clue. Does the light hitting their face match the lighting in the background?
Often, deepfakes have mismatched shadows or highlights, making the person look “pasted in.”
Look for blurry edges around the person, especially their hair or jawline. These can be tell-tale signs of manipulation. Even tiny details like earlobes or teeth can sometimes appear distorted or inconsistent.
Analyse Audio and Lip Sync
If it’s a video, the audio can be a huge giveaway. Listen carefully to the voice. Does it sound robotic, slightly off, or does the pitch change suddenly?
Sometimes, the background noise in the audio won’t match the visuals. For example, if someone is supposedly outside, but the audio is perfectly silent.
A major red flag is poor lip sync. Watch the person’s mouth as they speak. Do their lip movements perfectly match the words being said?
In many deepfakes, the lips might be slightly out of sync, or the mouth shape might not look natural for the sounds being made. It’s a subtle detail, but once you start looking for it, it becomes obvious.
Use Google Reverse Image Search
For images, a simple but powerful tool is Google Reverse Image Search. This can help you find out if an image has been used before, perhaps in a different context, or if it’s a known fake.
- Go to images.google.com.
- Click the camera icon (Search by image).
- Upload the image or paste the image URL.
Google will then show you where else that image appears online. If it pops up in many different places with unrelated stories, or if it’s linked to known hoaxes, that’s a strong sign it might be fake or misused.
Try Deepfake Detection Tools (InVID, FotoForensics)
Several free online tools can help analyse images and videos for signs of manipulation. They might sound technical, but they’re often quite user-friendly.
- InVID WeVerify: This is a browser extension (for Chrome and Firefox) designed for journalists, but anyone can use it. It helps you analyse videos from YouTube, Facebook, and other platforms. It can break a video into keyframes (individual pictures), perform reverse image searches on those frames, and even analyse metadata (information about the file). Look for inconsistencies in frames, or if keyframes show up in strange places online.
- FotoForensics: This website analyses images for manipulation using something called Error Level Analysis (ELA). Upload an image, and FotoForensics will show you a version where manipulated areas often stand out with different textures or brightness. Original parts of an image tend to have a consistent ELA pattern, while edited parts look different. It takes a little practice to interpret, but it can reveal hidden edits.
These tools won’t give you a definitive “fake” or “real” answer every time, but they provide strong clues that you can piece together.
Staying vigilant in the digital age is essential. While deepfakes are becoming more sophisticated, so are the methods to detect them. By combining careful observation with readily available tools, you can significantly improve your ability to identify fraudulent content and protect yourself from scams.
If you need more help and support, please check the links below
Our Digital Products
Our Digital Skills Training Services
Our Website design and maintenance Services
Visit our website for more Won Connect CIC