The cameo feature in Sora 2 enables users to embed their own likeness (face, voice, expressions) into generated videos. To enable cameo, a user records a short video + audio sample of themselves (or a consenting subject). The system uses that recording to build a representation of appearance, voice, facial dynamics, and expressive style. Once that representation exists, the user can allow or restrict its use in future generated content.
When a user permits their cameo, they effectively become co-owners of the generated content that incorporates their likeness. They retain rights to delete those videos or revoke future use. The app requires identity verification via that one-time sample to tie any future creations to the correct person, and ensures that the cameo cannot be used without consent. The system card also states that image uploads that include photorealistic people or video uploads are restricted initially, and more stringent safeguards apply for minors.
However, implementation of revocation, deletion of derivatives, and enforcement across reuse scenarios remains a complex challenge. Because users might remix or re-share videos, the platform must track provenance and prevent unauthorized reuse. Users will expect robust rights over how their likeness is used downstream. The cameo feature is designed to give individuals control and consent, but the effectiveness depends on system design and policy enforcement as Sora’s ecosystem scales.