(Credit: Kirk Wester / iStock Editorial / Getty Images Plus via Getty Images)
Don't miss out on our latest stories. Add PCMag as a preferred source on Google.
OpenAI may have deepened its feud with Hollywood with the launch of Sora 2, an AI video generator with hyper-realistic outputs.
The Hollywood Reporter (THR) spoke to multiple talent agencies, including WME, which say OpenAI either did not notify them of the product's launch or was "purposefully misleading" in the strength of the tool's content guardrails.
They argue OpenAI should require permission from a celebrities, studios, and animators, to use their characters in videos. WME represents actors like Ben Affleck, Christian Bale, Matt Damon, Denzel Washington, Jennifer Garner, and many other household names.
Pre-launch meetings with OpenAI executives, including COO Brad Lightcap, were "upbeat," the agents tell THR. The company seemed confident Sora 2 would protect their intellectual property and clients' likeness. Other OpenAI execs in the meetings allegedly included Sora product lead Rohan Sahai, media partnerships VP Varun Shetty, and talent partnerships lead Anna McKean.
While the tool launched with a ban on creating videos of public figures, some copyrighted content slipped through the cracks. Users could readily create scenes from recognizable movies, TV shows, and games, including Bob’s Burgers, SpongeBob SquarePants, Gravity Falls, Pokémon, Grand Theft Auto, and Red Dead Redemption, according to THR.
Some Sora 2 users also created mashups of well-known titles, like making Pokémon's Pikachu look like a character in The Lord of the Rings and Oppenheimer. Sora 2 users also quickly found an unexpected loophole—creating videos of dead celebrities, like Michael Jackson, which OpenAI banned a week later. The clips went viral, propelling Sora 2 to the top of the app store, although its ratings have since fallen to 2.8 on the Apple store.
The day of Sora 2's public release, WME released a memo saying it was opting out all of its clients from all AI-generated videos. Talent agency CAA, which represents the likes of Scarlett Johansson and Tom Hanks, also slammed the product, calling it "exploitation," CBNC reports. (Johansson sparred with OpenAI in 2024 after claiming it ripped off her voice.)
OpenAI CEO Sam Altman admitted the product's main goal was to make money, and to make people "smile." Hollywood is not.
"This was a very calculated set of moves [CEO Sam Altman] made," the WME exec tells THR, implying OpenAI knowingly released the product with some loopholes to earn attention on social media and drive ChatGPT subscriptions. "They knew exactly what they were doing when they released this without protections and guardrails."
Altman says he's hearing the opposite from many content creators who are excited about "this new kind of 'interactive fan fiction,'" in an Oct. 4 blog post. OpenAI still intends to give "rightsholders more granular control over generation of characters, similar to the opt-in model for likeness but with additional controls," but so far it has not disclosed any specific solutions of timelines.
WME is now exploring litigation against OpenAI, THR reports, though the legal guidelines for AI-generated content are still evolving, according to USC. Creatives had a win last month when a judge ordered Anthropic to pay $1.5 billion for training its models on copyrighted books.
Disclosure: Ziff Davis, PCMag's parent company, filed a lawsuit against OpenAI in April 2025, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.