Big Tech's deepfake detection efforts rely on C2PA labeling that barely works

Meta, Google, TikTok, and Microsoft have all joined the C2PA coalition for cryptographic media authentication, but adoption is inconsistent and ineffective. Meta only checks image metadata, not video. YouTube labels AI content without verifying authentic content. TikTok relies on creator self-labeling. A Verge analysis argues platforms are using C2PA as performative cover while simultaneously building and promoting generative AI tools. Instagram head Adam Mosseri acknowledged that "authenticity is becoming infinitely reproducible" but the platform continues going full speed on AI generation features.

View full digest for February 24, 2026