Can something be expected but huge? That’s what Disney‘s OpenAI deal is: the most inevitable bombshell around, but still a bombshell.
On Thursday the country’s biggest legacy entertainment company announced that it would be partnering with OpenAI to allow its characters to be toyed with in AI. Starting in the next few months, you’ll be able to play with Pixar, Marvel and Star Wars characters using Sora on Disney+, or on OpenAI’s own platforms. You want to crouch at a race starting-line next to a character from Cars, as a Disney sample showed? Your moment is here.
The inevitable part is that Disney CEO Bob Iger hinted several weeks ago something like this was coming, and most smart money had it involving OpenAI; the kind of social virality Iger suggested fit squarely with what OpenAI has been good at with the likes of Sora and ChatGPT.
Still, we didn’t know how soon this would happen. And the details weren’t known. One big detail: Disney is paying OpenAI $1 billion for an undisclosed stake in the company. Another detail: a lot of Disney characters will be available as part of the licensing component, but not such that you can see their face; the company is choosing to stay far from that SAG-AFTRA minefield (for now). Also, no voices, to avoid a voice-actor legal/ethical situation, though voice-mouth alignment in current models isn’t very good or convincing anyway.
But that still leaves a gigantic swath — so many animated characters, and presumably plenty of faceless ones like Darth Vader and Iron Man. (There will be content guardrails, Iger says.)
We’re about to be inundated on Disney+ with everyone making their own content — to type or likely soon speak a character into your own personalized story, the way you used to speak action figures into your own personalized story, except now it instantly becomes a video. Is this the world’s coolest Play-Doh or the world’s lamest memeslop? Just one of the key questions you’re forgiven for asking. Here are a bunch more, with some answers.
What’s motivating this move?
In a word, Netflix. That’s true historically — Netflix ran a 10K on streaming before Disney could lace up its sneakers, prompting a mad belated frenzy from the Burbank firm to get its own service up and running more than a decade later, in 2019. Bob Iger doesn’t want to see that happen again, so he’s getting out front on the next wave of digital entertainment — AI creations — before Netflix can add its own offering.
But it’s also true literally. Entertainment giants in the AI age will thrive if they excel on two criteria: tech and content. Netflix has the tech — no Hollywood platform has better algorithms, data or personalization engines. And Disney has the content — no Hollywood studio has more beloved characters and properties. What Netflix did earlier in the week was seek to bolster its content — by buying Warner Bros, it will have a library for training and AI character manipulations it couldn’t have built in 100 years (technically 102, given the age of WB). And now with the OpenAI deal, Disney is hoping it has the tech. Netflix. It’s always Netflix.
So how quickly does this thing go from faceless, voiceless characters to Captain America showing up at my holiday dinner table?
Iger is smart by keeping it away from the stuff people are most worried about: live actors getting manipulated to oblivion. There’s a Guild issue there, but there’s also an uncanny-valley issue there — are we really ready for Chris Evans to be walking and talking as we see fit? Animated characters are already stylized, so that valley is a lot shallower. “A Buzz Lightyear custom birthday card for their kid,” OpenAI leader Sam Altman said on CNBC Thursday. Which is pretty much the most benign, casualty-lite use case you could devise (unless you’re Hallmark)
But make no mistake: existing actors moving in lifelike ways (or synthetic actors doing the same) is where this is all going. Disney just needs to start making deals with unions and getting consumers comfortable first. Think of this as designing a car that looks like a carriage. Necessary, but temporary. Teslas will fill the road soon enough.
How should we think about OpenAI now?
How much time do you have? The most intriguing company in all of tech-dom is also the most frustrating to define. Founded as a do-gooder nonprofit a decade ago this month, morphed into a social-media ChatGPT “It” company three years ago this month, OpenAI is now something entirely uncategorizable. It has a knack for creating buzz thanks to Sam Altman’s evangelism; it has a tendency to be a lighting rod for everything we worry about thanks to Sam Altman’s evangelism. Its video tools are not seen as strong as Google’s Veo and its lack of app development is noticeable; the company waits for others to come along and build stuff for its models. Also, many, many people have left OpenAI because they worry about the company’s more … liberal attitude to safety.
The alignment with Disney, a company as American as they come, does a little burnishing. I mean can the place of Mickey Mouse and the Magic Kingdom really be leading us down an apocalyptic path? For all the ways Disney’s move sullies it with creatives in Hollywood, OpenAI’s new partner does the opposite — gives it massive cred in Silicon Valley. And, really, around the world. How un-American can Sam Altman be if Disney loves him?
That said, OpenAI is still building mysterious pins, promising AGI and releasing weirdly named Large Language Models, no matter how many Little Mermaid re-creations it enables. If a company this vexing didn’t exist it would be necessary to invent it.
Isn’t Disney suing Midjourney? And did they really just cease-and-desist Google on the same day they announced this OpenAI deal?
It can seem at first glance like the company is all over the place on AI; you can be forgiven for saying “pick a lane, Bob.” The truth is Disney has picked a lane: the express one, to the land of Generative AI.
What people thought back when Disney’s Midjourney lawsuit was filed this summer — that it really wanted human-led art — wasn’t really accurate. It wanted machine-generated art. It just wanted to control what was being fed into that machine. The alignment with screenwriters and craftspeople was just incidental.
By the way, if you think it’s only Disney, well. This will also open the door for others. Where one conglomerate plunges, others will soon tread. Expect Universal and Warners – despite their own Midjourney suits — to follow Disney into the pool. They may not have the same store of animated characters that Disney has. But they’ll find a way.
What does this mean for new movies and TV shows? Is Disney still as excited about that business?
Well of course if you ask them they will say yes. Iger himself told CNBC Thursday that “this does not in any way represent a threat to the creators at all. In fact, the opposite. I think it honors them and respects them.”
But he also said that when it comes to AI “you can’t do anything about it. No human generation has ever stood in the way of technological advance, and we don’t intend to try,” adding “if it’s going to happen, including disruption of our current business models, then we should get on board” – which hardly sounds like a man eager to keep television and film output at the current levels.
Part of what’s happening here is a concession — I wouldn’t call it a white flag, but I’d hardly call it a vote of confidence either. Iger is seeing just how much of the attention economy is being gobbled up by user-generated content on TikTok, YouTube, Insta et al. And rather than fight it, he’s leaning into it. He’s betting that instead of makeup tutorials and viral dances, AI now allows Disney can play the short-form user-generated game too, by making many of its properties available — that people looking to upload, create and share would rather do it with stuff they grew up loving, especially now that you can have those characters do pretty much anything you want.
And he’s betting that this is a majorly good use of the company’s time and resources — a form that drafts off what they have instead of the laborious, unproven expensive work of creating something new. That doesn’t mean Disney won’t be in the original content business, of course — if nothing else you need to refresh old characters or invent new ones for us to AI-play with it. But it does mean the focus is a little different; it does likely mean that in a world of limited dollars more of them are going to be rehashing what’s been made than making something new; it does mean that the artist-led work of film reboots and nostalgia plays could be given way to the masses-led work of rehashes and prompted entertainment.
Disney is still Disney. But as of today it’s a little less about finding new water sources and a little more about going back to the well.
Will this work?
The part nobody knows. Legit nobody. All this supply, in a very strange inversion of everything you learned in freshman Microeconomics, is coming before any evidence of demand. Will speaking characters into action be a novelty or a mainstay — a paradigmatic way of interacting with screens or just a passing diversion that will turn this deal cringe in short order? Will it be TikTok or MySpace? The bet feels like the former — people love these characters, and the stuff you could do with them is elastic and eye-opening. But nobody really knows. Iger’s billion will soon help us find out.
Read the full article here
