OpenAI has officially introduced Sora 2, its latest video generation model, which incorporates AI-generated audio synchronized with the created video. This release is accompanied by the launch of a new iOS application, also named Sora, which features a “Cameo” option allowing users to include themselves or friends in AI-generated videos. OpenAI asserts strong identity protection measures are in place to prevent unauthorized use of individuals’ likenesses in the videos.
Additionally, an application programming interface (API) for Sora 2 is planned, enabling third-party developers to integrate this model into their own video editing tools. An Android version of the application is also in development. Initially, users in the U.S. and Canada will have access, with a wider rollout planned for other regions. Sora 2 is offered for free, with usage limits for all users, while ChatGPT Pro subscribers receive access to a higher-quality model.
During a livestream, OpenAI’s team discussed the capabilities of Sora 2, emphasizing its advanced physics, realism, and control in generating various complex actions. The model aims to deliver a more cohesive experience by synchronizing dialogue and sound effects, with the Cameo feature allowing users to create personalized video interactions after a brief recording.
The new app encourages creativity over passive scrolling, allowing users to remix videos and personalize their feed. Additional safety measures for users under 18 include disabled infinite scrolling by default and stricter content safeguards. Privacy settings and parental controls are incorporated to enhance user safety.
OpenAI has outlined several safety measures regarding identity protection and content moderation. The company is also launching new features such as storyboards for video creation, with the API expected to be available shortly. CEO Sam Altman highlighted the importance of continuous user satisfaction and ethical considerations surrounding the service. While the capabilities of Sora 2 present opportunities for creativity, its success will also depend on how OpenAI manages user trust and safety priorities moving forward.
Source: https://venturebeat.com/ai/openai-debuts-sora-2-ai-video-generator-app-with-sound-and-self-insertion