OpenAI launches Sora, its first social video app powered by AI

OpenAI has officially launched Sora, a groundbreaking social application that integrates AI technology into video creation, marking the company's debut in the short-form video arena. Currently available on iPhones in the U.S. and Canada, Sora operates on an invite-only basis and delivers an experience reminiscent of TikTok, where users can create, share, and remix video clips. What sets Sora apart from conventional video platforms is its focus on AI-generated content. Central to the app is a unique feature called "cameos," which permits users to upload a single video and audio recording of themselves. This verified likeness can then be utilized to create short clips, either by the user or by their friends, while granting full control over access and deletion rights. OpenAI emphasizes that this co-ownership model empowers users to dictate how their likeness is utilized. Additionally, Sora includes a "Remix" function designed to promote user engagement with existing clips and trending topics. Currently, videos generated through the app are capped at a duration of 10 seconds. The system curates recommendations based on user activity, geographic location, engagement levels, and optionally, interactions with ChatGPT. In response to concerns surrounding screen time and content safety, Sora incorporates parental controls, allowing guardians to set limitations on scrolling and personalization features. To mitigate misuse risks, OpenAI has implemented several safeguards. For instance, public figures can only appear in Sora videos if they personally upload their cameo, and explicit or extreme content is automatically filtered out. Users retain the ability to revoke permissions at any time, although there are ongoing worries regarding the potential for non-consensual or misleading AI-generated material. Sora is currently free to use, with plans for monetization through paid options for additional video generations during peak demand. Early adopters will also have the opportunity to invite up to four friends, although an Android version has yet to be confirmed. Behind Sora's innovation is Sora 2, OpenAI’s advanced video generation model. This upgraded system enhances physical realism and storytelling across various shots, delivering more natural movements and improved audio-video synchronization. Moreover, ChatGPT Pro subscribers can access Sora 2 directly without needing an invite. Looking ahead, OpenAI has hinted at future enhancements, including a Pro version, web access, and a developer API. Employees who participated in the internal pilot have described Sora as a potential game-changer for video content creation, likening its impact to that of ChatGPT in the realm of conversational AI.