Scammers are placing Taylor Swift into fake interviews she never gave, then running those clips as TikTok ads to steal personal data from fans. Researchers who mapped these campaigns found the videos use AI-manipulated footage to make it appear Swift - and other celebrities like Rihanna - are endorsing products or inviting viewers to claim prizes. The clips are effective enough that scammers keep running them, meaning the economics work.
This is the problem driving Swift's reported effort to trademark her own likeness - a legal move that would give her clearer standing to sue over unauthorized use of her image, whether AI-generated or not.
How the Scam Actually Works
The technique doesn't require expensive production tools. Scammers take real footage of celebrity interviews or public appearances and run it through AI video software to modify the audio and, in some cases, lip movements to match new dialogue. The result doesn't need to be flawless - it needs to be convincing enough for someone scrolling through TikTok to pause and click.
Tools that produce this kind of face animation have dropped dramatically in cost over the past two years. What once required professional video production can now be done with a consumer software subscription. Products like D-ID have made realistic AI-generated face animation accessible to anyone with a credit card, which also means the volume of potential scam content has outpaced what platform moderation can realistically catch.
Once a viewer clicks through, they land on a phishing site or a form asking for their email, phone number, or payment details in exchange for a "prize." One clip can run as a paid ad to millions of users before TikTok's moderation flags it.
Why Trademarking Your Face Is a Complicated Fix
Swift's trademark strategy reflects a real gap in current law. The right of publicity - the legal concept preventing unauthorized commercial use of someone's name or image - varies by state and is difficult to enforce across international platforms operating under different jurisdictions.
Trademarking a name or likeness creates a different legal lever. Instead of arguing that publicity rights were violated, a trademark holder can argue consumer confusion - that fake content misleads viewers about the source or endorsement of a product. It's a more reliable claim in court.
The limits are significant though: trademark law is designed to protect consumers from being misled about goods and services, not to give individuals comprehensive control over their image. Courts have been reluctant to extend trademark protection far into personal identity. And enforcement against anonymous overseas scammers is nearly impossible regardless of which IP rights you hold.
The more durable solution is platform-level detection. TikTok already bans synthetic media that deceives users, but catching AI-manipulated video at scale is technically hard when the manipulation is subtle - a clip that modifies lip movements while preserving original lighting and camera angles can pass basic review systems looking for obvious artifacts. For users, the practical rule is simple: any ad featuring a celebrity personally inviting you to claim something for free is fake.