OpenAI's new video tool can use copyrighted content by default: report

OpenAI's new AI video generator, Sora 2, will use copyrighted material from television and movie studios by default unless rights holders explicitly opt out, as per reports from Reuters and The Wall Street Journal.

The policy means content from films and shows could appear in AI-generated videos created by users unless copyright owners contact OpenAI to remove their material, states the reports.

According to the Reuters report, at least one major studio, Disney, has already chosen to opt out. The approach mirrors OpenAI's existing policy for image generation and follows the company's earlier lobbying for AI training on copyrighted content to be considered "fair use" under US law, adds the reports.

Alongside the model, OpenAI has launched a social iOS app called Sora where users can create, share, and remix AI-generated videos. The app includes a 'cameo' feature that lets people insert realistic, AI-generated versions of themselves into videos after verifying their identity. Videos in the app can be up to 10 seconds long and include synchronised audio and sound effects.

As per its official blog, OpenAI says Sora 2 represents a significant technical step forward, with improved physics simulation and the ability to handle complex actions and multiple shots. The company also emphasised safety measures, such as identity verification and user controls over AI-generated likenesses.

The Sora app is being rolled out initially in the US and Canada through an invite-only system. OpenAI plans to make the video generator available for free at launch, though it may introduce fees later depending on demand and computing capacity.