OpenAI tightens controls on copyrighted characters

OpenAI’s newly released video generation model, Sora 2, has quickly gone viral for producing hyper-realistic clips inspired by popular cartoons and video games — from South Park to Pokémon.

But the company is now moving to give copyright holders more control over how their characters appear in AI-generated content.

OpenAI CEO Sam Altman announced that the company will implement stricter policies to address copyright concerns after Sora 2’s debut last week.

“We will give rightsholders more granular control over generation of characters,” Altman wrote in a blog post on Friday, describing the update as similar to OpenAI’s opt-in model for likenesses, but with additional controls for copyrighted material.

The move follows a series of copyright lawsuits against OpenAI, including a major case filed by The New York Times, as well as growing concerns over the use of intellectual property in AI-generated content.

According to the Wall Street Journal, OpenAI plans to allow studios and copyright owners to opt out of having their content replicated in videos made with Sora 2.

Sora 2, launched on October 1, includes a TikTok-style app that lets users insert themselves into AI-generated scenes.

Although the tool has begun rejecting prompts involving Disney or Marvel characters, users have still managed to produce and share videos depicting Japanese and U.S. franchise characters, such as Pokémon’s Pikachu, Super Mario, and Sonic the Hedgehog.

The debate over Sora 2 underscores the growing tension between AI innovation and copyright protection, as content creators and major studios push back against the unlicensed use of their intellectual property in generative models.