AI is remaking entertainment. Are you ready?
The Hollywood strikes miss the forest for the trees: the real disruption will come from AI-generated content created by someone sitting in front of her laptop.
Introduction
While actors, directors, and studios fret about box office numbers, streaming service competition, and shifting audience preferences, artificial intellignce looms in the background. AI technology is progressing at a breakneck pace, and it is poised to revolutionize not just filmmaking, but the very roles that humans play in creating movies. Many dismiss this idea. Many are convinced that complex human emotion can’t be algorithmically replicated. Yet, this skepticism may prove to be unwarranted. A lot of people fail to grasp how rapidly AI’s capabilities are improving. We are quickly approaching a time where the creation of a movie is directed not by a studio and the personnel it hires, but rather by an individual sitting in front of her laptop.
This presents a profound shift in entertainment production, away from both Hollywood studios and the writers and actors they employ, and toward the lone individual. Picture someone sitting at her computer, interrogating an AI about a movie idea that she has. Through an iterative process the person and the machine synthesize a new movie from opaque algorithms. Maybe this sounds fanciful to you, but the technology to do this is already here. The technology is admittedly rudimentary: algorithmically-generated movies can’t yet approach the quality of conventionally-generated movies. But the technology is quickly becoming more powerful, and today’s artificial intelligence is the worst you’ll ever use1.
AI technology is rapidly traversing a J-curve: its capabilities are improving at a faster rate over time. This kind of superexponential growth is not intuitive to many people. Most people look at a given piece of technology and have a static view of it and its capabilities. In this mental model of technology, the computer that you use today is as powerful and as capable as the computer you used last year, or the computer you will use next year. And while this static view of technology is sometimes adequate, when it comes to artificial intelligence, it is utterly misleading. The rate at which the technology is improving is itself increasing. That second derivative—the rate of change of change—is what accounts for AI’s j-curve. And that is what skeptics miss about AI technology and movies.
So, whether you’re a filmmaker, an actor, or just an interested observer, it’s time to pay attention.
The Current State of AI in Filmmaking
The incursion of AI into the world of filmmaking is not a distant dream. It’s already here. Take scriptwriting, for instance. Algorithms can now generate entire scripts based on provided prompts or even past works of acclaimed writers. Casting is another domain where AI is making headway. Facial recognition algorithms can suggest actors who would be a perfect fit for a role, thus reducing the dependency on casting directors to some extent. In post-production, AI can now handle tasks like color grading, sound mixing, and even editing, accelerating the production timeline considerably.
Seven years ago, well before today’s generative AI hype, IBM used its Watson AI to create a trailer for the movie Morgan.
Even more intriguing are smaller projects entirely generated by AI algorithms, from plot to final cut. While these may not have reached commercial success, they serve as proof of concept and a glimpse into what the future could hold. You can see many of these movies via LucidBox. Here’s a short film about creativity in the age of AI:
To the skeptics, the current applications of AI in filmmaking might appear amateurish and unrefined. However, this perspective may stem from a static view of technology. What is often ignored is that AI is on an upward trajectory, both in terms of quality and versatility. The technology of today is not stagnant; it’s continuously evolving. Algorithms get refined, data sets expand, and the computational power backing these systems multiplies. The nascency of AI-generated art should not be a justification for dismissing its potential; rather, it should be an indicator of the massive untapped possibilities.
So, even as AI-generated works may lag behind human-created films in nuanced storytelling or emotional depth, it is crucial to recognize that the gap is closing. One only has to compare AI-generated movies or media from a year ago to those created more recently to witness significant advancements.
The view that AI will never catch up with human creativity underestimates not only the capability of the technology but also the very nature of innovation. We should not be asking whether AI can replicate human creativity but when and how it will revolutionize it.
The Mathematics of AI Progress
When we talk about super-exponential growth, it’s easy to think of it as just another buzzword. But in the context of AI, it’s a powerful concept that governs the rapid advancements we’re seeing. To put it simply, super-exponential growth means that the rate of improvement isn’t constant—it’s perpetually increasing. While linear growth adds a constant amount at regular intervals—think of it like walking at a steady pace—super-exponential growth is more like a snowball rolling down a hill, gathering more snow and speed as it goes along.
You may have heard of the “J-curve” in various contexts, but here, it serves as a visual representation of how the rate of AI’s capabilities is accelerating over time. The initial part of the curve is flat, representing a period where progress seems slow and incremental. However, the curve eventually takes a sharp upward turn, illustrating the point where each improvement is significantly bigger than the last.
Let’s move from the abstract to the concrete. Think about AI’s role in facial recognition technology. A decade ago, the technology was rudimentary, often struggling with basic tasks like identifying faces under different lighting conditions. But today, not only can it identify faces, it can also gauge emotions, age, and even who the individual resembles, offering precise casting suggestions in seconds. Apple introduced Face ID six years ago.
Consider another example in scriptwriting. Initial algorithms were capable of generating short, simple dialogues but struggled with coherent long-form stories. Fast-forward to now, and we have algorithms that can generate complex narratives with plot twists and character arcs. They are not just imitating human scripts but are creating original content, often in a fraction of the time a human writer would need.
Even in post-production, AI tools that initially could only manage basic cuts and transitions can now analyze entire movies, suggest edits, optimize sound quality, and more. These are not marginal improvements; they represent leaps in capabilities.
What ties these examples together is the underlying exponential growth. Each new iteration of these AI tools is not just a bit better than the last; it’s often significantly more capable, enabled by more sophisticated algorithms, more extensive training data, and faster computational power.
This is not to say that we’ll see an Oscar-winning AI-generated movie tomorrow. However, the trend lines are clear. The pace of innovation is not constant—it’s accelerating. And given this trajectory, the window for dismissing AI’s role in filmmaking as “amateurish” is rapidly closing. The technology is catching up, and it’s doing so faster than most of us realize.
Empirical Evidence
While theories and projections about AI’s capabilities are interesting, nothing speaks louder than empirical evidence. An ever-growing repository of AI-generated media is available for anyone interested in gauging the technology’s progress. For this discussion, let’s focus on a particular Netflix-style collection that curates various AI-generated works called LucidBox.
What follows is a number of examples of AI-generated short movies collected on LucidBox. These should give you a taste of what is out there.
Here’s a spoof of Wes Anderson and Star Wars:
Here’s a realistic AI-generated simulation of a screenwriter figuring out the story he’s trying to tell:
Here’s animation done in the style of Pixar:
If you dive into this collection and look at projects that were produced six months ago, a year ago, and those that are recent, you’ll notice something fascinating: the quality isn’t just improving. It’s doing so at a noticeably rapid pace. Six months ago, you might find that AI-generated characters had wooden expressions or that the narratives lacked nuance. However, more recent projects reveal characters with emotional range and stories with complex plot lines, even integrating elements like humor and suspense that are traditionally difficult for machines to replicate.
This isn’t a claim based on a few outliers. Across the board, the increase in quality and complexity over a short time is observable. The improvements in visual effects, character development, and even soundtracks are more than incremental—they often leap over what used to be significant barriers.
Here’s YouTuber Cleo Abram intereviewing Spotify CEO Daniel Ek and musician Grimes about AI’s role in creating music:
She and Grimes demonstrate a tool which allows anyone to copy Grimes’ singing voice. Imagine someone using AI to write dialogue, and then record it in Travis Bickle’s voice.
Stuff like this will keep the movie studios and their lawyers up at night, but it’s also inevitable.
These advancements aren’t just cosmetic. They hold the potential to fundamentally change the way stories are told and acted out. Imagine AI being able to adapt a script in real-time based on audience reactions or fine-tuning a character’s expressions to the minutest detail, surpassing even what human actors can emote.
While it’s tempting to dismiss AI-generated media as a passing fad or a series of interesting experiements, the data shows otherwise. There’s a trajectory here, marked by rapid, substantial improvements. Given that this is happening in multiple aspects of filmmaking—scripting, acting, editing—it’s tough to argue that this is isolated or coincidental. The empirical evidence suggests that AI is not only improving but is doing so in a manner that will inevitably bring it into direct competition with traditional filmmaking roles. Ignoring this reality may be convenient today, but could prove costly in the very near future.
The Threat to Traditional Filmmaking
As AI continues to encroach upon various aspects of filmmaking, the question naturally arises: what becomes of human actors, directors, and other creative professionals? The answer isn’t necessarily dystopian. Humans will likely move into roles that AI can’t easily replicate, at least for the foreseeable future. This could mean more focus on abstract creative direction, nuanced storytelling, or specialized acting skills that are challenging for AI to emulate. However, there’s no denying that the job descriptions in Hollywood and worldwide will undergo transformations.
AI’s rapid advancements also herald a change in the business models of the film industry. Independent filmmakers could harness AI tools to produce high-quality movies at a fraction of the traditional cost, significantly lowering the barriers to entry. On the flip side, established studios might find themselves needing to adapt or become obsolete. They could either become early adopters, integrating AI into their workflows, or risk becoming irrelevant as smaller players deliver comparable quality at a lower price point.
The infusion of AI into filmmakking isn’t just a technological or economic issue. It has ethical and artistic implications as well. For example, as AI begins to generate realistic human-like characters, issues around representation and cultural sensitivity could arise. Furthermore, the concept of ‘originality’ in art will be challenged. Is a compelling story less valuable if it’s generated by an algorithm rather than a human? These are questions that the industry will need to grapple with sooner rather than later.
The growth trajectory of AI in filmmaking is a reality that neither skeptics nor enthusiasts can afford to ignore. For those in the industry, now is the time for critical engagement with these technologies. Brushing them off as amateurish or treating them as merely tools for human facilitation misses the broader picture. We’re on the cusp of a revolution that could redefine art, entertainment, and the very concept of human creativity.
The signs are clear, and the empirical data is compelling. Whether one sees AI’s role in filmmaking as a threat or an opportunity, indifference is no longer a viable stance. The industry, as well as audiences, must prepare for a future where AI isn’t just an accessory in the creative process but a major player, capable of both imitating and innovating. And as we move forward, it’s crucial to approach this not as a looming calamity but as an evolving landscape, rich with new challenges and unprecdented possibilities.
The notion that “today’s AI is the worst you’ll ever use” is, while profound and accurate, hard to place. I don’t know who first came up with this construction, and a Google search has proven fruitless. But it’s not a construction I invented.