What Are Deepfakes?
Imagine watching a video of a famous person saying something shocking or hearing a phone call from a friend that just doesn’t sound right. With today’s technology, it’s now possible to create videos, images, or audio where people appear to say or do things they never actually did. These are called deepfakes.
“Deepfake” is a word that combines “deep learning,” which is a type of AI, and “fake” content.. It means digital content, usually videos or audio, that has been cleverly altered or entirely created by computers to look and sound real.
How Do Deepfakes Work?
Deepfakes are created with powerful computers and extensive data, including images and videos of a person’s face or recordings of their voice. This data is used by specialised computer programs to “learn” how a person looks, talks, or moves. The more data and computer power you have, the more real the deepfake looks. For example, with enough videos of a celebrity, a deepfake creator can make a believable video of that celebrity saying something they never actually said.
Why Are Deepfakes Important?
Deepfakes can be fun and even helpful, like creating realistic movie effects or bringing historical figures to life in documentaries. But they come with serious risks:
- Deepfakes can generate fake news or videos that are meant to trick people, which makes it harder to identify what’s true online.
- Fraud and Scams: Criminals might use deepfaked audio or video to impersonate someone and trick people out of money or secrets.
- Privacy and Reputation Damage: Deepfakes can be used for bullying, blackmail, or to harm someone’s reputation.
How Can You Spot a Deepfake?
While deepfakes are getting better, there are still clues:
Unnatural Facial Movements
- Eyes: Deepfakes often have trouble making flashing patterns that look real. Keep an eye out for folks whose eyes don’t move smoothly or who don’t blink often enough.
- Mouth and Lips: The mouth movements might not match the spoken words exactly, or the lips might seem out of sync with the voice.
- Facial Expressions: Emotions may not look right—smiles might appear forced, or expressions may change too quickly or awkwardly.
Strange Visual Artefacts
- Skin Texture: The skin tone on the face may not match the neck, hands, or background, or the face may look excessively smooth or hazy.
- Edges and Outlines: Look for fuzzy or flickering edges around the face, especially when the person moves their head or their hands pass in front of their face.
- Glitches: You might spot brief glitches or “melting” effects, especially if the video quality drops.
Inconsistent Lighting and Shadows
- The lighting on the face might not match the rest of the scene. For instance, the shadows can fall the wrong way, or the face might look brighter or darker than the background.
Awkward Head and Body Movements
- The head may move unnaturally or fail to align with the body.
- The person’s body language or gestures might look stiff, robotic, or out of place.
Audio Clues
- Voice Quality: The voice may sound flat, robotic, or slightly “off” compared to known recordings of the person.
- Lip Syncing: Watch for words that don’t line up perfectly with lip movements.
- Background Sounds:Sometimes background noise doesn’t match the setting or cuts off abruptly.
Odd Context or Behaviour
- The person says or does something completely out of character, controversial, or shocking.
- The setting, clothing, or actions don’t make sense for the person or the situation.
Technical “Tells”
- Video Quality: Deepfakes may be lower in quality than real videos, with less detail or strange compression effects.
- Eye Reflection: Eye reflections may not accurately reflect the environment or light sources.
Check the Source
- Is the video or audio coming from a trusted, official channel? If not, be extra cautious.
- Look for reputable news outlets or the person’s own official accounts to confirm if the content is real.
If you see something suspicious, check whether trustworthy news sources are reporting it as well, or use online deepfake detection tools.
How to Protect Yourself
You don’t need to know much about technology to stay secure. Deepfakes might be convincing, but there are things everyone can do to reduce the likelihood of being targeted. This is how you can stay safe:
Be Sceptical of Surprising Content
- If you see a video or hear an audio clip that seems shocking, outrageous, or entirely out of character for the person shown, pause before believing or sharing it.
- Ask yourself: “Does this seem real? Would this person really say or do that?”
Verify Before Sharing
- Don’t rush to share content just because it’s dramatic or fascinating.
- Check if trustworthy news sources or official channels cover the same story.
- Do a quick online search to see if fact-checkers or reputable websites have flagged the content as fake.
Look for Signs of Manipulation
- Pay attention to how the person’s face moves. Do the eyeballs blink on their own? Does the mouth say what the words say?
- Check for strange lighting, awkward head movements, or “glitches” around the face.
- Listen carefully, does the voice sound flat, robotic, or a little off?
Protect Your Personal Content
- Limit the amount of personal photos, videos, and audio clips you share on public social media, especially if your profile is open to everyone.
- Consider making your social media accounts private and only sharing content with people you trust.
Use Strong Privacy Settings
- Review and update the privacy settings on your social media accounts to determine who can access your personal information and posts.
- Be careful about accepting friend requests or follows from somebody you don’t know.
Report Suspicious Content
- Most social media platforms and websites allow you to report content that seems fake or harmful.
- If you see a deepfake being used to harm someone, report it to the platform and consider alerting the person being targeted.
Educate Yourself and Others
- Keep up with the latest news regarding deepfake technology and how to spot manipulated media.
- Share what you learn with friends, family, and colleagues so everyone can be more cautious together.
Enable Two-Factor Authentication
- Two-factor authentication can help keep hackers out of your accounts.
Use Trusted Tools
- Some browsers and apps offer plug-ins or tools that can help detect manipulated media, and consider using them for extra peace of mind.
The Takeaway
Deepfakes show us how powerful and creative technology can be, but they also make us think about what we see and share online. As deepfakes become more common, learning to question and verify digital content is one of the best ways to protect ourselves and our communities.
Stay curious, stay cautious, and always check twice before believing something you see on the internet!