
Photo Credit: Rihanna for Fenty Beauty / CC by 3.0
Over the past month, TikTok has experienced a rise in videos of AI-generated celebrities, including Taylor Swift, Rihanna, and Kim Kardashian, encouraging users to sign up for fraudulent rewards programs or other “malicious services.” Fortunately, the synthetic celebs aren’t very convincing and feature a lot of uncanny facial movements and asynchronous audio—but that doesn’t stop scammers from baiting the hook.
According to Copyleaks, besides just scammy ads, the deepfakes also include hyper-sexualized ads promoting nonconsensual face and body swapping, or digital “undressing.” Examples of the clips include a digitally altered Taylor Swift advertising a so-called “TikTok Pay” feature, and a similarly altered Rihanna pushing a views-for-rewards program.
Many of the clips use TikTok branding and add textured filters, seemingly to bypass the platform’s AI-detection tools. If users click on the videos, they will be redirected to third-party sites that solicit their personally identifiable information.
The timing is notable, since Taylor Swift recently filed trademarks to protect her voice and likeness from AI misuse following repeated incidents with deepfakes. But the problem is widespread and shows no signs of slowing down.
Meta was recently hit with a class-action lawsuit filed on behalf of the Consumer Federation of America with allegations that the company profited from Facebook advertising scams like the ones proliferating on TikTok. Similarly, the FTC noted that scam ads are one of the top methods used to push shopping-related scams on social media, and AI is only making that problem worse.