OpenAI will shut down its Sora tool

OpenAI plans to shut down its Sora text-to-video tool, a stunning move that comes three months after Walt Disney Co. pledged to invest $1 billion in the artificial intelligence company and allow the use of dozens of beloved characters.

The San Francisco-based company did not disclose why it was shutting down the tool or the timeline for its phaseout. In a post Tuesday on the Sora account on X, the company said it knew the news was “disappointing.”

“To everyone who created with Sora, shared it, and built community around it: thank you,” the post said.

Open AI’s pivot comes as the company was engaged in discussions with Disney to formalize their arrangement — but no deal had been reached, according to a source familiar with the matter who was not authorized to comment.

Although Disney had pledged to make the huge investment, the company had not yet made any payments to OpenAI, this person said. OpenAI had not paid any fees to license Disney characters.

A Disney spokesperson said in a statement that the company respected OpenAI’s decision to shift its priorities away from video generation.

“We appreciate the constructive collaboration between our teams and what we learned from it, and we will continue to engage with AI platforms to find new ways to meet fans where they are while responsibly embracing new technologies that respect IP and the rights of creators,” the spokesperson said.

The emergence of Sora had roiled Hollywood, particularly as AI and compensation for actors’ likeness and voice became a central issue in the 2023 strike.

Performers guild SAG-AFTRA had said at the time of the Disney-OpenAI announcement that it would “closely monitor the deal and its implementation to ensure compliance with our contracts and with applicable laws protecting image, voice, and likeness.”

OpenAI first previewed Sora in 2024, and the realism of the tool’s AI-generated videos grabbed audiences at a time when competing video generation apps struggled.

The text-to-video platform enabled users to create short videos with different styles, voices and dedicated features such as “storyboard,” which enabled users to weave together prompts to make longer videos with consistent characters — something that wasn’t possible before.

In September, OpenAI launched Sora as a dedicated app to create and share AI-generated videos with friends, which many viewed as a social networking app modeled after TikTok.

The app’s remix feature enabled users to superimpose the likeness of their friends or celebrities into existing AI-generated video or create new ones. Sam Altman, the chief executive of OpenAI, encouraged users to slap his likeness onto AI-generated scenes and other pop-culture videos.

The lax approach to copyright allowed the re-creation of dead celebrities and copyrighted characters from titles including WWE and South Park, which OpenAI said it would allow on its platform unless the celebrities opt out.

Despite hitting 1 million downloads in a week, the app lost its sheen, as regular users found little everyday use for a dedicated AI video app. As legal challenges mounted, Sora also strengthened its copyright guardrails and “content violation” warnings became a routine part of denying user requests.

But the AI space has become increasingly crowded. OpenAI’s smaller rival Anthropic has gained ground by offering its AI coding services to enterprises rather than just to consumers. Its Claude tool has become especially popular for coding tasks.

Since Sora’s release, competitors such as Google Veo and Bytedance’s Seedance also have rushed into the AI video generation market.

Times staff writer Meg James contributed to this report.