OpenAI Is Preparing to Launch a Social App for AI-Generated Videos (www.wired.com)

OpenAI is preparing a standalone app built around its next-generation video model Sora 2 that delivers vertical, swipeable short-form AI videos in a TikTok-like feed, WIRED reports. The app limits clips to about 10 seconds, offers a For-You recommendation page, like/comment/remix controls, and does not allow uploading user photos or videos—everything is generated by the model. It includes an identity verification feature: verified users can opt to let others use their likeness in generated clips, and people receive notifications when their likeness is used (even for drafts). OpenAI has piloted the app internally with positive feedback so far. For the AI/ML community this signals a major push to normalize synthetic video as a mainstream content medium and to productize generative video UX (short-form feed + remixing). Technical and policy implications are significant: Sora 2 still struggles with realistic physics and complex action scenes, and OpenAI’s filters already block many generations for copyright or safety reasons amid ongoing lawsuits (e.g., The New York Times). The app will compete directly with Meta and Google’s recent AI-video offerings and raises fresh challenges around consent, deepfakes, moderation, age restrictions and liability for likeness use—areas that will demand both model-level safeguards and platform governance.