OpenAI is rolling out new measures to give copyright holders more granular control over how their characters are used in its AI video tool, Sora. The move aims to address mounting concerns from studios, creators, and legal voices over unauthorized use of intellectual property in AI-generated content.
Sora, OpenAI’s text-to-video app, has already generated short videos featuring recognizable intellectual property, including characters from SpongeBob SquarePants, Pokémon, South Park, and more, without prior permission from rights holders. Critics have warned of copyright infringement and appropriation risks.
Sam Altman, OpenAI’s CEO, confirmed that feedback from rights holders shaped the decision. The company is exploring a revenue-sharing model for permissible use of characters. He acknowledged that some “edge case” generations may slip through guardrails initially but pledged ongoing improvements.
Under the new approach, content owners could gain the ability to block or restrict the usage of their characters in Sora videos, akin to how users currently have opt-in control over their own likenesses. However, OpenAI clarified that it does not yet support a blanket opt-out for all use. Instead, creators will gain more precise control, allowing them to restrict specific uses, content types, or modifications.
Varun Shetty, OpenAI’s head of media partnerships, said they’ll work closely with studios and media companies to enforce character blocks when requested. The company has already received takedown and opt-out demands and is building tools to support those workflows.
“Rightsholders… are very excited for this new kind of ‘interactive fan fiction’ and think this new kind of engagement will accrue a lot of value to them, but want the ability to specify how their characters can be used (including not at all),” Altman justified Sora.
Despite changes, several studios have already acted: Disney has opted out of allowing its characters to be used in Sora. Others are reportedly evaluating their position.