TikTok tests tool that detects use of your face in AI-generated videos

TikTok is testing a tool that can identify when a person's face appears in videos generated or manipulated by artificial intelligence. The idea is simple: the platform compares the faces that pop up in content against a user's registered likeness and alerts them when it finds suspicious matches. That way, anyone whose image is used without permission in a deepfake can be warned and request the video's removal faster than is possible today.

The news comes at a delicate moment: deepfakes have become cheap and easy to produce, and cases of scams, bullying, and fake intimate content have exploded worldwide, including in Brazil—one of the biggest social-media markets. Platforms now face pressure from governments and users to act. Detecting people's faces is a different layer of protection from just labeling AI-generated videos, because it directly defends each user's identity on the network.

For those learning AI, this news highlights an important side of the technology: it's not enough to create; you need to protect. Understanding how facial-similarity detection works and the limits of these systems helps build a critical view of the risks of generative AI. What's more, it's a practical relief for anyone who posts photos and videos online—especially Brazilian content creators, who earn their living from their own likeness and can be targets of malicious impersonation.

A practical example of impact: imagine someone uses an app to put your face on a fake video begging your followers for money—a scam that already happens in Brazil. With the tool, TikTok could detect the facial match, alert you, and limit the video's reach before it fools many people. The same logic extends to protecting artists, journalists, and even students who are victims of embarrassing mashups in school or work groups.

Practical advice: while the tool hasn't rolled out to everyone, activate the protections already available—review your account privacy, report suspicious videos, and be wary of money requests made 'on video,' even with a familiar face and voice. It's also worth talking with your family about deepfakes and agreeing on a safety word. Keep an eye on the next steps: if the test works, other networks, like Instagram and Kwai, are likely to adopt similar solutions soon.