OpenAI Launches Sora; Copyright Opt-Out Raises Concerns

On Sept. 30, 2025, OpenAI launched Sora 2, a next‑generation video-and-audio generation model, and an invite-only Sora iOS social app in the U.S. and Canada that produces short (about 10‑second) AI-generated clips in a TikTok‑style feed. Sora 2 adds synchronized audio, improved controllability, multi‑shot continuity and stronger adherence to physical realism across multiple visual styles, and OpenAI positions it as a rival to Google’s Veo‑3. The app’s Cameo system lets people upload short video and audio samples to create verified digital likenesses for themselves or friends; creators are recognized as co‑owners of remixes and can revoke access. OpenAI built in safety and identity safeguards — including liveness checks, bans on generating public figures without opt‑in, prohibitions on adult/extreme content, invite‑only rollout, usage caps, and possible paid extras — and said it will offer a Sora API and an Android release. OpenAI’s copyright approach defaults to allowing copyrighted works to appear in generations unless rights holders opt out (Disney has reportedly opted out), a stance that has already prompted legal and industry concerns and has drawn commercial interest from investors and traders.