On Tuesday, September 23, 2025, Sam Altman, CEO of Openai Inc. during a media tour of Stargate AI Data Center in Abilene, Texas, USA.
Kyle Grillo | Bloomberg | Getty Images
Openai is debuting what it bills as the most advanced video generation model, along with a new iOS app designed to provide artificial intelligence-powered video creation to everyday users.
The SORA 2.0 model is built as a research preview in previous versions released this year. Openai said the new release represents a major leap in physical realism, audio-video synchronization, and multi-shot storytelling.
A new app, also known as SORA, is only an invitation from the start, allowing users to create, remix, and cameos with AI-generated videos using text or images with built-in controls for safety, likeness and origin.
The deployment has already raised concerns about intellectual property regulations.
The Wall Street Journal reported Monday that it was notifying studios and talent agencies that their copyrighted material could be reflected in SORA-generated content unless Openai explicitly opts out.
Openai says that its approach is based on existing content policies used in image generation tools.
The company views this style of engagement, in which users remix or refer to the familiar fictional world, as a form of fan expression and creative play.
Executives say they have a strong user interest in using AI to interact with their beloved stories and characters, and Sora believes they are introducing new ways for people to engage in those fictional universes.
Openai also draws a clear line between copyrighted IP and individual portraits.
You may see characters and visual styles from popular franchises unless they are blocked, but you cannot generate individuals with prompts or images unless you have submitted a verified cameo.
This policy aims to give people complete control over how, when, and how AI will appear in generated content.
The app will be rolled out on an invitation-only system in the US and Canada on Tuesday. The pro version, web access and developer APIs are all coming soon.
Watch: Humanity launches the latest AI model, Claude Sonnet 4.5
