How Deepfakes Work & How to Protect Yourself Against them.

What is a Deepfake?

A deepfake is a video, image, or audio created by AI to imitate a real person.?

Over the last couple of years, AI services and products have reached mainstream usage, resulting in billions of pieces of AI generated content being created.

Using deep learning algorithms like GANs (Generative Adversarial Networks), computers can now generate shockingly realistic fake media.?

Think AI-generated Taylor Swift singing your favorite song—but it’s not Taylor Swift! And she’s not singing your favorite song. In fact, she’s not singing at all.?

With AI making deepfakes easier than ever to generate, the risk they pose to everyday people is on the rise.

From AI phone scams to video call scams costing $25M USD – bad actors are using this technology, and they’ll be applying it more frequently over coming years. Once you understand how they work, mitigating the risk of becoming a deepfake target is easy.

How Are Deepfakes Made?

1. Collecting Data: AI needs video data to learn a person’s facial expressions. This can be anything from:Videos on social mediaPublic videos like interviews?Videos collected via data leaksVideos collected via hacks or stolen account access.

Naturally, celebrities and public figures are the most comment target for Deepfakes – such as Content Creator Gibi, Mr Beast, Taylor Swift, Scarlett Johansson and many more.

1. Training the AI: The AI watches hours of footage, learning to mimic facial movements and voice patterns. After training, it can recreate your face and even voice – saying, doing, and expressing using your face.?

1. Creating the Fake: Once trained, usually the face will be used as an overlay on another video… of something else happening. To the untrained eye, these can sometimes be nearly indistinguishable from a video of something else happening.?

Why You Should Care

With the rise of celebrity deepfakes like Jenna Ortega, Pokimane, or Ariana Grande, it’s getting harder to spot what’s real. AI tools are becoming more and more accessible, meaning more people have the power to create deepfakes. This means that normal people (not just celebrities) face an increasing risk of having their identity stolen and used in malicious ways.?

To protect yourself against the consequences – fake scandals, identify theft, or personal blackmail, you need to take control of your digital footprint, and manage it diligently. This means deleting content that doesn’t need to be online, and applying the strictest privacy settings to every account you own.

What you can do about it:

Concerned about the amount of content you’ve posted, and the potential risk it opens you up to? Here are a few things you can do right now to reduce risk

1. Audit and clean up your digital footprint. The more content you share online, the more material bad actors have to create deepfakes. Limit public sharing and consider purging your public-facing content with an app like ours – redact.dev.?
2. Tighten your social media and privacy settings. Set your accounts to provide, enable two-factor authentication, and regularly review public data exposures with tools like Google Alerts or reverse-image search.
3. Be cautious with video and voice data sharing. Phone and video scams can be hugely damaging – be wary of unsolicited calls and video calls, avoid sharing voice memos, video messages and live streams if possible.
4. Stay updates on deepfake trends and scams. Follow cybersecurity blogs and newsletters (like this one!). Consider making a redact.dev account and opting into email updates.?

Take control with redact.dev. Our tool lets you clean up multiple platforms quickly and easily – avoid deepfakes, targeted attacks and make sure your digital footprint is well managed.