OpenAI says it’s working with actors to crack down on celebrity deepfakes in Sora

OpenAI said Monday it will make more efforts to block users of its AI-powered video creation app Sora From creating clips in the likenesses of actors and other celebrities after actor Bryan Cranston and the union that represents film and television actors raised concerns that deep fake videos were being created without the consent of the performers.

Actor Bryan Cranston, the Screen Actors Guild-American Federation of Television and Radio Artists (SAG-AFTRA) and several talent agencies said they have reached an agreement with the maker of ChatGPT over the use of celebrity photos in Sora. Subscriber statement It highlights the intense conflict between AI companies and rights holders such as celebrity estates, film studios, and talent agencies – and how productive AI technology continues to undermine reality for all of us.

Sora, a new sister app to ChatGPT, allows users Create and share AI-generated videos. It was launched to great fanfare three weeks ago, and AI enthusiasts have been searching for it Invitation codes. But Sora is unique among AI-powered video generators and social media apps; It lets you use other people’s recorded likenesses to put into almost any AI-powered video. It was, at best, bizarre and funny, and at worst, An endless scroll of deepfakes Which is almost indistinguishable from reality.

Cranston noticed Sora users using his image when the app launched, and the Breaking Bad actor alerted his union. The new agreement with the actors’ union and talent agencies confirms that celebrities will have to choose to make their images available for placement in the AI-generated video. OpenAI said in the statement that it “reinforced barriers around sound repetition and similarity” and “regretted these unintended generations.”

OpenAI has guardrails to prevent the creation of videos of known people: You declined my request to request a video of Taylor Swift on stageFor example. But these guardrails aren’t perfect, as we saw last week with a growing trend of people creating videos featuring the Rev. Martin Luther King Jr. These clips ranged from bizarre deepfake of the civil rights leader rapping and wrestling in WWE to explicit racist content.

Don’t miss any of our unbiased technical content and lab reviews. Add CNET As Google’s preferred source.

A deluge of “disrespectful images,” as OpenAI called it In a statement On Friday, that’s part of the reason the company is pausing the ability to create videos featuring King.

His daughter Bernice A. asked. King, last week, publicly asked people to stop sending her father AI-generated videos. She was Echoing Comedian Robin Williams’ daughter, Zelda, called these types of AI videos “horrible.”

OpenAI said it “believes that public figures and their families should ultimately have control over how their photos are used” and that “authorized representatives” of public figures and their estates can request that their photos not be included in Sora. In this case, King’s estate is responsible for choosing how his image is used.

This isn’t the first time OpenAI has relied on others to make these calls. Prior to Sora’s launch, the company informed a number of Hollywood-adjacent talent agencies that it would have to opt out of including their intellectual property in Sora. But this initial approach did not comply with decades of copyright law — typically, companies need to license protected content before using it — and OpenAI. She reversed her position After a few days. It’s one example of how AI companies and innovators work Conflict over copyrightIncluding through High-profile lawsuits.

(Disclosure: Ziff Davis, CNET’s parent company, in April filed a lawsuit against OpenAI, alleging that it infringed Ziff Davis’s copyrights in training and operating its AI systems.)