Amazon’s Next Big Bet: An AI Animation Tool for Hollywood
At a moment when rivals are racing to perfect digital humans, Amazon appears to be quietly sketching out a different future for film.
I recently attended a private screening of the new 20th Century Studios/Hulu original film Swiped, about the creation of the dating app Bumble and its founder, Whitney Wolfe Herd. Despite some mixed reviews, I had a great time, as did most of the audience, based on the reactions in the theater. And the question-and-answer session after the film was enlightening, as the film’s star and producer, Lily James (Downton Abbey, Baby Driver, Cinderella), director and writer Rachel Lee Goldenberg, and executive producer Gala Gordon all gave deeper insight into the production.
What I wasn’t expecting was that, after leaving the theater, there would be a crowd of James fans waiting outside the relatively obscure Directors Guild of America (DGA) theater in Manhattan. (Fun fact: Director Christopher Nolan has just been elected president of the DGA.) The fans were waiting for a chance to meet James, and perhaps get an autograph, or share a brief gaze with their film and television star. They wanted to connect with the person who had embodied so many meaningful stories over the years and delivered a kind of magic that they had kept with them.
And that got me thinking again about AI and, more specifically, generative AI films.
The Gravity of Stars
Watching the rapid development of the generative AI video landscape, it seems the ultimate goal, the brass ring, the “championship,” if you will, is: Who can produce consistent, truly realistic human characters in AI first? That’s what all these monthly iterative breakthroughs in generative AI video are moving toward.
Special effects? This is where AI is already making its biggest Hollywood mark, first by erasing jobs and forcing traditional VFX professionals to discover and harness AI tools that can enhance their existing skill sets. Realistic cityscapes and landscapes? Gen-AI video is already doing an increasingly impressive job at replicating the look of real places and imaginary worlds with convincing fidelity.
Background actors? Well, if we’re talking zombie hordes, and your run-of-the-mill bustling Times Square New York background crowd scene, again, it seems clear that we’re close to seeing believable AI outputs of those scenes. And that may crater the already low-paid entry path for many actors who rely on background work to get a foot in the door for bigger roles and industry connections.
What if no one wants AI humans to replace human actors? And by “no one,” I mean even the movie studios, the supposed target customers of these billions of dollars in funding that fuels many generative AI video startups.
The aforementioned AI video replacements are coming, probably sooner than we expect. But none of the AI video players like Runway, OpenAI (Sora), LumaLabs, Google (Veo 3), and others will be satisfied until they can produce consistent artificial humans that look the same from scene to scene. Furthermore, they want these AI actors to mimic the physics of how real humans interact with each other, deliver dialogue in a natural way, and transcend the uncanny valley of artificiality that triggers a slight revulsion in audiences.
But I’ve been thinking: What if no one wants AI humans to replace human actors? And by “no one,” I mean even the movie studios, the supposed target customers of these billions of dollars in funding that fuels many generative AI video startups. In the often-whispered, sinisterly framed narratives of Hollywood business skeptics, movie studios are portrayed as only caring about the bottom line, with the ideal arrangement being one in which they control actors like marionettes. And that may be true to some extent. However, movie studios sell several tiers of products: mass-market entertainment catering to the lowest common denominator, family films and animation, mid-budget genre films, and award-winning prestige films that help differentiate them from their competitors, stoking a greater sense of wonder and mystery around the alchemy of producing visual stories.
This is what most tech CEOs are getting wrong about this race to produce artificial human lead actors for AI-generated films. Hollywood isn’t just about the cookie-cutter assembly line stuff that gets churned out. Hollywood isn’t just looking for a more efficient and lower-cost way to keep the sausage factory part of its business going. A large part of what makes Hollywood what it is resides in the human stars that fans seek a connection with.
The fans are craving what Horton and Wohl, in their 1956 paper “Mass Communication and Para-Social Interaction,” called parasocial relationships. These one-sided relationships fans have with their stars are reinforced by media stories detailing the stars’ lives, audio, video, in-person interviews, as well as ongoing news about their family or personal projects. This parasocial bond solidifies what was initially just a brief fascination with an on-screen performance into an enduring connection with another human who fans know lives and breathes in the same world as they do. People relate to people. Will this change? Can humans, at scale, form long-lasting parasocial relationships with people (not avatars or creatures, but one-to-one replicas of mundane humans) who don’t exist in the tangible world? Maybe. But nothing, not even AI therapists and their odd mental grip on some, has convinced me that people will stop wanting to be connected to real people…even if just parasocially.
Drawn by Machines
For the above reasons, I believe the leading area of theatrical and TV entertainment that has the most immediate chance of being wholesale replaced by AI video in the near future is animation. This is one place where being a kind of artificial person is celebrated and illuminated. The largest consumers of animation are children (adult animation fans notwithstanding). And this is not a picky group when it comes to film and TV.
Partly because they are still developing into whoever they will be, they generally don’t care if the animation from Nickelodeon and Disney is made in Los Angeles or South Korea. They just want to be entertained, and provenance is irrelevant to this cohort. And because it’s far easier (for now) to produce human-comparable animation via generative AI than live video, it’s a no-brainer that animation will, in most cases, become mostly AI-generated or largely AI-assisted.
There will also be genre films in the realm of horror and perhaps comedy that audiences may accept as fully or mostly AI-generated productions. Think about it, who are the modern horror icons we know the best: Jason Voorhees (mask), Michael Myers (mask), Pennywise the Clown (mask), as well as an assortment of creatures and dolls that don’t require human-level verisimilitude. Phantasmagoric imagery and costume are the stock-in-trade of horror, making it perfect for AI video.
Nevertheless, on a larger scale, animation will be the target for AI-generated video. So who understands this and is making strategic moves in that direction? I have an answer.
A Secret In Seattle
Other than Netflix, Amazon probably has the best data on what humans really want when it comes to video entertainment. That doesn’t mean they can always deliver (hey there, War of the Worlds), but Amazon Prime Video, tied into the Amazon store filled with synergistic merch, electronics, and even food, gives Amazon unique insight into what may be next in entertainment.
That’s why I found it extremely notable that, while everyone else is focused on creating AI humans to populate future films, Amazon Studios is instead quietly building an AI-powered software suite geared toward animation.
Earlier this month, Prime Video & Amazon MGM Studios listed job openings for a Senior Design Technologist and a Principal UX Designer to work on creating an as-yet-unnamed AI software suite for movie and TV industry creators. And while we don’t have the name, we have a rough idea of what the AI software suite will do based on the job listing.
Described as “a startup within Amazon,” the person filling these roles will develop “a next-generation, AI-powered creative suite that blends art and science to make the impossible feel effortless…”
Described as “a startup within Amazon,” the person filling these roles will develop “a next-generation, AI-powered creative suite that blends art and science to make the impossible feel effortless — uniting ideation, storyboarding, character design, animation, scene composition, localization, real-time iteration, and more in a unified workspace that amplifies, not replaces, human creativity.” Another version of the same passage describes this as all being done, “all while preserving the magic of storytelling.”
Further hints at how the AI software suite will function are revealed in additional position details: “We’re especially interested in candidates who’ve built tools for animation or worked closely with creative roles (e.g., technical artist, animator, etc.) and who are comfortable bridging traditional digital content creation pipelines and tools (e.g., Maya, Blender, Unreal, etc.) with modern AI-native creative solutions.”
Obviously, there’s no guarantee that this “startup within Amazon” will be solely devoted to Amazon. Plausible adjacencies within the all-in-one software suite could include AI VFX tools and even AI-powered game building. Still, the comparisons to traditional 3D modeling software, as well as the “magic of storytelling” phrase, indicate that Amazon may be cooking up what amounts to an AI animation master tool for rapidly building animated films and TV shows that will be designed to be used in-house and possibly licensed to other studios.
If it is what it looks like it is, that is, AI animation, I think this is the right direction for AI in Hollywood for those who want to own the near-term future of the technology and how it becomes integrated into film and TV. Of course, Amazon has proven to be more adept at product sales and technology development rather than groundbreaking storytelling. So building this AI master storytelling tool will be just the first part. The real test will be how creatives use it when it’s fully baked and ready to start producing work that people actually want to watch. After all, that’s the point of the entire Hollywood game: Can you weave a good yarn, no matter what the tools are, AI or not.