Over the last several months, media executives have become increasingly willing to at least acknowledge the inevitability of artificial intelligence when it comes to content production. Not only has Netflix co-CEO Ted Sarandos mentioned FX shots achieved with AI, but Warner Bros. Discovery CEO David Zaslav (who may become part of Netflix at some point based on M&A talks) has praised the new AI-fueled version of WBD’s “Wizard of Oz” at Sphere in Las Vegas. In fact, both he and Sphere owner James Dolan even made AI-generated cameos in that production. In August, AMC Networks CEO Kristin Dolan trumpeted the company’s partnership with AI video firm Runway, which is also working with Lionsgate and others. And around the same time, Amazon’s former Prime Video U.S. chief Albert Cheng – formerly of Disney and a longtime media exec – became Head of AI Studios, with Amazon Head of Prime Video & Amazon MGM Studios Mike Hopkins (also a traditional media veteran) charging him to “focus on empowering the very best filmmakers and showrunners to do their most innovative work in AI.” Those are only a few of the developments from both the traditional and “Silicon Valley” side of the media equation. There will be more. Everyone, it seems, has crossed the AI Rubicon.
But despite all of the above, this week’s watershed deal between Disney and OpenAI was different as it goes well beyond the tinkering and exploring phase and bleeds into sustainable integration. Disney’s $1 billion investment in OpenAI includes what would have been unthinkable even six months ago: Disney will license its beloved characters to OpenAI’s Sora AI video engine so that anyone can create short-form videos featuring anything from Mickey Mouse to Darth Vader, who we admit deserve a good lightsaber battle in the extremely near future. Guardrails will supposedly restrict usage to avoid problematic character depictions, with Disney CEO Bob Iger saying the deal will instead “thoughtfully and responsibly extend the reach of our storytelling.” With that promise comes a key risk that Disney will gradually lose control of how the public sees its IP, much of which remains part of ongoing franchises. And Disney is already suing Midjourney and other companies for failing to stop users from spoofing copyrighted characters. Only a day before it announced the OpenAI investment, Disney also sent a cease-and-desist letter to Google claiming that its AI training methods and large language models infringe on its copyrights on a “massive scale” while “leveraging its dominance in generative AI and across multiple other markets to make its infringing AI services as widely available as possible,” per Deadline.
The irony, of course, is that Sora has been trying to do that for months, especially since launching a social media app on Sept. 30 aimed at creating a TikTok-like community of AI short-form video creators. The app has yet to become the next big thing, partly because OpenAI must block so many copyright- or likeness-infringing prompts from users. But deals like the one with Disney could open the floodgates as users find they can do almost anything with established characters (no drugs, sex, alcohol, or interactions with non-Disney IP, of course). But such guardrails were meant to be broken, and seasoned prompt engineers have become increasingly adept at tricking AI models into breaking their own rules. It’s almost certain that unauthorized usage of Disney characters will slip through the Sora cracks, and it’s unclear how Disney and OpenAI will respond when that happens. Because those risks abound, it’s likely that Iger and other Disney execs made the calculation that those breaches will happen anyway – so why not at least be in the room where it happens rather than suing from outside the tent? The big question is whether other traditional media companies with considerable IP like WBD will follow suit. And Disney probably understands that chief IP rival WBD especially won’t be able to make any bold moves into AI at least until after a clear winner emerges from the M&A drama between Paramount Skydance and Netflix. And even then, it will be at least 18 months to two years before the new regime takes over and can set any long-term AI strategy for its valuable library and IP. By then, Disney will have a big head start; its OpenAI deal is only exclusive for three years and could serve as a training ground for wider licensing. Disney also plans to lean in either way, even planning to feature some user-created Sora videos on Disney+. We’re heading into extremely uncharted territory – and no one knows for sure what monsters lurk in the shadows.
Next? It’s important to note that the Disney-OpenAI deal is happening against an uncertain regulatory and financial backdrop. This week, President Trump signed an executive order banning states from heavily regulating AI after California, Colorado, New York, and others stepped in to set up broader guardrails to promote transparency and enact safety, consumer, and anti-discrimination protections. The fear is that such restrictions will hamstring the industry as it competes to maintain U.S. dominance over China (Ironically, Trump simultaneously lifted some chip restrictions this week on Nvidia, which can now sell its advanced H200 chips in China.) The schism between several states and the federal government on this likely will grow and lead to protracted court battles. Case in point: New York Gov. Kathy Hochul signed that state’s new AI bill the same day as Trump signed his executive order, and New York and other states are reportedly coordinating legal action to challenge the order in a fight that could end up in the U.S. Supreme Court. As that all unfolds, Hollywood will continue to work through its own AI debate, with Disney and others doing deals as unions representing actors, writers, animators, and others raise objections and point to clauses in recent unions contracts designed to protect creators who fear IP character usage only opens the door to a looser attitude that could lead to AI-written screenplays, AI-produced movies, and even AI-created actors like the infamous Tilly Norwood. Meanwhile, the overall AI machine chugs along with seemingly reckless abandon as OpenAI also released its new ChatGPT 5.2 model this week as a response to Google’s Gemini 3 launch last month. The competition is fierce. It’s studio vs. studio, AI model vs. AI model, the U.S. vs. China, the suits vs. the artists, and the states vs. the feds. And we’re only getting started. Perhaps we should just hand all of this over to Skynet and hope for the best.
