Imagine a tech mogul outmaneuvering an entire industry, turning Hollywood's stars and stories into tools for his empire—without so much as a polite ask. That's the jaw-dropping tale of Sam Altman and OpenAI's Sora, and trust me, it's a plot twist that's got everyone talking.
Picture this: Late September, mere days before OpenAI unveiled its brand-new social media app tied to its cutting-edge video generator, Sora. Rumors were flying about shifts that could seriously challenge Hollywood's grip on its creations. One big talent agency, catching wind of the buzz through industry grapevines, was baffled—why hadn't Sam Altman's powerhouse startup looped in their top brass? So, they took the initiative and made contact.
The initial discussions between the agency and OpenAI were anything but smooth. According to an executive from the agency who was right in the thick of it, OpenAI's reps were 'intentionally deceptive.' The tech team's vibe was super positive, hyping up a system where everyone would opt in to protect the agency's clients' intellectual property and their likenesses—think faces and voices—from being exploited.
But here's where it gets controversial... OpenAI's crew, including big names like CFO Brad Lightcap, Sora's product leader Rohan Sahai, media partnerships VP Varun Shetty, and talent partnerships head Anna McKean, assured Hollywood that the upgraded safeguards on Sora 2.0 would make everyone thrilled. Spoiler alert: That promise didn't hold up. As the agency exec shared, 'We compared notes with other groups in similar chats and realized we were all getting mixed messages.'
Key inconsistencies popped up, especially around how likenesses (like a celebrity's face) versus intellectual property (the broader creative works) would be handled. Some agency leaders were promised an opt-in for everything, while others heard the reverse—or nothing at all. OpenAI's communication to Hollywood felt scattered and unreliable.
Take WME, for instance, home to A-listers like Matthew McConaughey, Michael B. Jordan, and Ryan Reynolds. A partner there was bluntly informed that their clients would need to proactively tell OpenAI to exclude their faces and voices from the platform. He was shocked. Recalling the conversation, he said, 'Picture an agent phoning a client today to pitch signing up for Sora—it's a surefire way to get fired!' He pushed back hard on OpenAI, emphasizing, 'We've got a massive roster of talent. If every single one opts out, what's the point of your whole project?'
Then, on September 29—just a day before Sora 2 launched—this WME partner got word that likenesses wouldn't be used without explicit permission. It was a partial victory for the agencies when OpenAI finally released the social media app, letting users whip up content with Hollywood's blockbuster properties owned by major studios. But here's the catch: Studios have to manually flag each property they want off-limits, marking a bold new step in Silicon Valley's push into entertainment.
And this is the part most people miss... By October 3, Altman declared a shift to something resembling an opt-in system—but not quite. The strategy had already paid off. Early on, Sora churned out clips featuring famous movies, shows, and games like Bob's Burgers, SpongeBob SquarePants, Gravity Falls, Pokémon, Grand Theft Auto, and Red Dead Redemption. Now, it's the top free app on the App Store.
'It was a meticulously planned series of actions,' the agency exec reflected. 'They launched without any real protections, knowing full well what they were doing.' Negotiations, now involving lawyers, are heating up, and lawsuits aren't off the table.
As OpenAI charges forward, Hollywood is playing catch-up. Even the usually quiet Motion Picture Association—the industry's chief lobbying group that's mostly sat out the AI debates so far—felt compelled to publicly call out the company. Agencies like CAA and UTA joined in, with WME's digital strategy chief Chris Jacquemin sending a memo to agents, stating he'd informed OpenAI that none of their clients would be part of Sora's latest edition.
Behind the scenes at studios and agencies, there's a mix of confusion and anger over the opt-out process. Insiders say there's no streamlined system; instead, it's a clunky link for reporting violations, like flagging issues one by one—a painfully slow method. The agency exec noted that OpenAI lacked dedicated staff for Sora's safeguards before launch (though they have some now), and many in Hollywood are now wary of engaging with them at all.
Some believe that just talking to OpenAI puts the burden on them to specify what IP can't be touched, forever. In a stern letter to OpenAI at the end of September, Disney made it clear: 'We're not obligated to opt out to protect our copyright rights.'
This summer, Disney, Universal, and Warner Bros. Discovery kicked off what could be a wave of legal battles by suing Midjourney for letting users generate content with their copyrighted characters.
'They're flipping copyright upside down,' explained Rob Rosenberg, a partner at Moses Singer and former Showtime Networks executive VP. 'It's a phony deal where they can proceed unless you object—and if you don't, it's on you.'
For years, OpenAI's approach under veterans like Shetty and McKean has been straightforward: Seek forgiveness after the fact, not permission upfront. It's the easiest way to monetize Sora. ChatGPT has exploded, raking in about $1 billion monthly from around 700 million weekly active users, according to reports from The Information. But Sora's path is less clear. AI video tools thrive on producing recognizable content—think how Midjourney added protections after a lawsuit only to drop them when usage plummeted. Altman is loudly promoting Sora's potential, which hinges on studios licensing their material.
In a blog post, Altman wrote, 'Many rights holders are buzzing about this 'interactive fan fiction' and see it as a way to build huge value through fan engagement.' Shetty agreed, noting, 'It opens doors for creators to forge deeper bonds with audiences.'
Hollywood is in a grueling war of attrition against a flush AI sector—and it's falling behind, reminiscent of the early internet days when piracy ran rampant due to slow responses. Their main bargaining chip? Future partnerships. Yet OpenAI approaches as a rival, not a friend.
'How do you expect collaboration when you've already burned the bridge?' the WME partner challenged OpenAI's team.
Inside agencies, there's frustration that studios aren't pushing back harder against Silicon Valley's creep into Hollywood. AI firms have long trained on pirated movies and shows from the web. Creators often lack control since their work is typically done for hire, meaning they don't own the rights.
But wait, here's a controversial angle to ponder: What if studios had sued sooner? Could they have forced AI companies to wipe their training data, like authors did with Amazon's Anthropic? Currently, studios are obsessed with safeguarding their IP over exploring AI collaborations. It's a tricky tightrope—some might bypass deals and launch their own apps, charging subscriptions for fan-made Disney content, while others eye partnerships. And this is where agencies and studios might clash.
'As the WME partner put it, 'If they litigate, they shut the door on future alliances.'
A rival agency exec was blunter: 'Our stars demand protection. This highlights bigger problems.'
To wrap this up, what do you think? Is OpenAI's 'ask forgiveness' strategy genius innovation or reckless theft? Should Hollywood sue now and risk missing out on lucrative deals, or partner up despite the risks? Share your take in the comments—do you side with the tech giants or the creatives? Let's debate!
This piece originally ran in the October 15 edition of The Hollywood Reporter magazine. Click here to subscribe.
