TikTok has identified more than 3 billion AI-generated videos on its platform using a recently tested deepfake detection feature. This tool is currently available to a select group of US creators who opt into the program.
To participate, creators must verify their identities through Jumio, which requires a live selfie scan and government ID check. TikTok states it does not retain any ID documents or store facial data after verification.
Once verified, the tool automatically scans for AI content that mimics the creator's likeness. If suspicious videos are detected, creators can report them for TikTok's review and possible removal. This move follows YouTube's 2025 launch of its own similar detection system, initially for partner creators, later expanded to journalists and civic figures.
TikTok uses a combination of the C2PA Content Credentials standard, supported by Adobe and Microsoft, along with proprietary watermarking and detection methods to identify AI videos. The C2PA protocol addresses content provenance issues that many blockchain projects also seek to solve.
Jumio, the verification service provider for TikTok's new tool, also manages identity verification for TikTok Shop and other platforms, ensuring verified digital identities are increasingly integrated into online services.
material is informational and not financial advice
