OpenAI has launched Sora 2, an advanced AI video generation tool that allows users to insert themselves into generated videos complete with synchronized sound. The update aims to make video creation more personal and accessible through a new app. Experts highlight both entertainment potential and deepfake risks.
OpenAI unveiled Sora 2 in October 2025, building on its previous text-to-video model. The key innovation is a feature enabling users to upload a personal image, which the AI then integrates into custom video scenes. For instance, a user could appear as a character in a fantastical adventure, with the video including generated audio like dialogue or ambient sounds.
According to Ars Technica, the tool processes user inputs via a dedicated mobile app, where prompts describe scenes, and the AI handles the rest. 'Sora 2 democratizes video production by letting anyone star in AI-crafted stories,' an OpenAI spokesperson stated. The release follows Sora's initial debut in 2024, with improvements in realism and length—videos now up to 60 seconds.
Wired reports that the app targets entertainment applications, such as short films or social media content, but raises alarms over deepfake implications. 'This could blur lines between real and fabricated media,' noted AI ethicist Dr. Elena Vasquez. OpenAI has implemented safeguards, including watermarks on outputs and restrictions on sensitive prompts, to mitigate misuse.
The launch occurred amid growing scrutiny of generative AI. Background context includes OpenAI's shift toward consumer tools after partnerships with media firms. No major contradictions appear between sources; both confirm the October rollout and core features. Implications include boosted creativity for creators, though regulators may push for stricter controls on synthetic media.
User testing began immediately post-launch, with early feedback praising ease of use but calling for better customization options.