How AI Will Change Hollywood Forever
For the past 25 years, I have been fortunate to be involved in several AI breakthroughs while also using the other half of my brain to create films and stage productions. This dual experience gives me a unique perspective on this topic, as I have a deep appreciation for both fields and can see a clearly defined and predictable outcome.
So, let me ask the big question: What happens when anyone can make a world-class film for just $120?
Generative AI such as Sora and other models will change the way movies are made. Period. And this change will be rapid and dramatic…but it isn’t what you are thinking. Everyone is asking how this technology will reduce costs for studios. Such as the headline from this recent story in Indiewire “Will AI Make Hollywood Production Cheaper”: https://www.indiewire.com/news/business/will-ai-make-hollywood-production-cheaper-1235006290/
But this is the wrong question entirely. What GenAI will do is drive the cost of quality content creation to near zero, potentially obsoleting Hollywood.
To understand this we have to ask…
Why does Hollywood exist?
Hollywood was formed in the late 19th and early 20th centuries as a center for the emerging motion picture industry, which was attracted by the favorable climate and natural scenery of Southern California. At the time, this was a brand-new technology and entertainment medium. And entirely new industry.
To make motion pictures, several elements were needed, such as:
- A camera that could capture and record moving images on film, such as the kinetograph invented by Thomas Edison.
- A projector that could display the recorded images on a screen, such as the vitascope developed by Charles Francis Jenkins and Thomas Armat.
- A film studio that could provide the necessary facilities (like large sound stages), sets, equipment, and personnel for producing movies, such as Universal Studios.
- A distribution network that could deliver the movies to theaters and audiences, such as the Motion Picture Patents Company or the Paramount Pictures Corporation.
- A star system that could create and promote popular actors and actresses, such as Mary Pickford, Charlie Chaplin, or Rudolph Valentino.
In other words, Hollywood exists because today it takes specialized people and equipment to make a quality film. So you need a system that employs these experts, buys and stores the equipment, and even builds huge sound stages to film on. This has been required for 100 years…until now.
Important Industry Statistics
· Total worldwide spend on film entertainment tops $100B/yr (including theater, streaming and physical distribution)
· Some 500,000 people work in the film industry in the US.
· Hourly wages across all job types averages $42/hr (about $85,000/yr)
· The best paid actors make between $25M and $100M/yr
· The top grossing lead actors have stared in movies bringing in as much as $6B in their career
· The top grossing career film director topped $10B (Steven Spielberg)
· All total over the past 100 years the US film industry has generated several trillion dollars of revenue from all sources.
This is a massive industry that is about to face severe disruption.
History shows us that technological advancements can profoundly disrupt industries by drastically lowering the cost of content creation and distribution, thus altering the revenue model. A prime example is the newspaper industry.
Newspaper circulation and revenue grew steadily for a century, peaking in 2005 at $60B, including advertising revenue and circulation. Today, those numbers have plummeted to less than $20B combined. The internet was the disruptor, turning everyone into a reporter and allowing anyone to distribute news at zero cost. This drove the cost of content creation and distribution to nearly zero, causing complete market disruption.
When cost dynamics change by 1000X or more, disruption is inevitable. But what happens when the cost is reduced by 10,000X, 100,000X, or even a million times? The incumbent industry faces a rude awakening. Even if it tries to embrace the new technology, the sheer volume of low to no-cost content overwhelms all but the most uniquely positioned entities. In the newspaper industry, only a few, like The New York Times and The Wall Street Journal, managed to survive by offering uniquely curated content. Perhaps one in ten newspapers survived, but fewer than five have built a sustainable business — less than 1%.
The film industry is poised for a similar transformation, and the consequences will be profound.
Take a look at the chart below documenting the demise of newspapers and what I see is a trajectory that Hollywood will soon follow:
What is the budget for a feature film?
A typical feature film costs about $1M/minute of run time. Often 300 to 2000 people are employed on one film project. While the breakdown of costs in a Hollywood production can vary widely depending on factors such as the scale of the film, genre, talent involved, and distribution strategy, here’s a rough average breakdown of costs based on industry standards:
1. **Above-the-line costs**: Typically, above-the-line costs can range from 20% to 40% of the total budget. This includes expenses for talent such as actors, directors, producers, and screenwriters. However, for major blockbusters or films with A-list stars, this percentage can be significantly higher.
2. **Below-the-line costs**: Below-the-line costs generally make up the majority of the budget and can range from 60% to 80% of the total budget. This category encompasses expenses related to production, post-production, crew salaries, equipment rental, set construction, costumes, props, special effects, and more.
3. **Marketing and distribution costs**: Marketing and distribution costs can vary depending on the distribution strategy and marketing campaign for the film. On average, marketing and distribution expenses may account for around 10% to 20% of the total budget. However, for major studio releases, especially big-budget films, marketing costs can sometimes exceed the production budget.
4. **Overhead and administrative costs**: Overhead and administrative costs are generally a smaller percentage of the total budget, typically ranging from 5% to 10%. These expenses cover general operational costs such as office rent, utilities, salaries for administrative staff, and other day-to-day expenses associated with running a production company.
5. **Contingency**: It’s common for productions to allocate a portion of the budget, usually around 5% to 10%, as contingency to cover unexpected expenses or cost overruns that may arise during filming.
Again, these percentages are rough averages and can vary depending on the specific circumstances of each production. For example, independent films or lower-budget projects may allocate a larger portion of the budget to below-the-line costs, while big-budget blockbusters may prioritize above-the-line expenses for high-profile talent and marketing efforts. Additionally, distribution strategies, such as theatrical releases versus streaming platforms, can also impact the allocation of funds.
What job types are involved in a feature film?
In addition to actors onscreen, there are numerous key positions and roles involved in making a Hollywood film. Here are some of the major ones:
1. **Director**: The director is responsible for overseeing the creative aspects of the film, including guiding the actors’ performances, making decisions about cinematography, and ensuring that the overall vision of the film is realized.
2. **Producer**: Producers are involved in the financial and organizational aspects of filmmaking. They secure funding, hire key personnel, and oversee the overall production process.
3. **Screenwriter**: The screenwriter is responsible for writing the script or screenplay, which serves as the blueprint for the film. They develop the story, create dialogue, and structure the narrative.
4. **Cinematographer/Director of Photography (DP)**: The cinematographer or DP is responsible for the visual aspects of the film. They work closely with the director to determine the look and feel of each shot, including lighting, framing, and camera movement.
5. **Production Designer**: The production designer is responsible for the overall visual aesthetic of the film. They work closely with the director and cinematographer to design sets, choose locations, and select props and costumes that help tell the story.
6. **Editor**: The editor is responsible for assembling the raw footage into a coherent and engaging final product. They work closely with the director to shape the story, pace the film, and create a seamless flow of images and sounds.
7. **Composer**: The composer creates the musical score or soundtrack for the film. They work closely with the director to develop music that enhances the emotional impact of the story and contributes to the overall atmosphere of the film.
8. **Sound Designer/Sound Mixer**: The sound designer is responsible for creating and editing sound effects, while the sound mixer is responsible for balancing and mixing all of the audio elements to create the final sound mix for the film.
9. **Visual Effects Supervisor**: In films with visual effects, the visual effects supervisor oversees the creation of digital effects and ensures that they seamlessly integrate with the live-action footage.
10. **Production Manager/Line Producer**: These individuals are responsible for managing the day-to-day operations of the production, including scheduling, budgeting, and coordinating logistics.
11. **Assistant Director (AD)**: The assistant director assists the director on set, helping to coordinate the cast and crew, maintain the shooting schedule, and ensure that filming runs smoothly.
12. **Stunt Coordinator**: In films with stunts or action sequences, the stunt coordinator is responsible for designing and executing these sequences safely.
13. **Costume Department**
13. **Additional Saff**: Such as food, payroll, handlers etc.
How good will the Gen AI thing be?
Generative AI is a branch of artificial intelligence that focuses on creating new content, such as images, text, audio, or video, based on some input data or parameters. Generative AI technologies, such as generative adversarial networks (GANs), variational autoencoders (VAEs), and transformers, have shown impressive results in generating realistic and diverse content, ranging from faces, landscapes, and artworks, to music, stories, and speeches.
In this article, we will explore how generative AI can impact the film industry, both in terms of production and consumption, by enabling new ways of creating, editing, and distributing movies. We will also discuss some of the challenges and opportunities that generative AI poses for filmmakers, actors, and audiences, as well as the ethical and legal implications of using these technologies.
It is already exceptional and getting better by the week. Take a look at this short montage of Sora generated scenes from early 2024: https://youtu.be/4GQ1axkFHz4?si=4a9ii3TeIy-KWiO9
No cameras. No actors. No locations. No drones. Just typing text.
Do I actually need a costume department or costumes when I can Outfit anyone?
https://www.youtube.com/watch?v=jnNHcLdoxNk
Or I can just anu=imate any phot and bring it to life like Portrait Alive:
https://www.youtube.com/watch?v=VlJ71kzcn9Y
One of the primary applications of generative AI in filmmaking is to replace or enhance the current methods of computer-generated imagery (CGI) and visual effects (VFX), which are typically expensive, time-consuming, and labor-intensive. Generative AI models have the potential to create high-quality and photorealistic images and videos from a few examples, sketches, or descriptions, without extensive manual work or specialized hardware.
For instance, Generative Adversarial Networks (GANs) can create synthetic backgrounds, props, or characters that blend seamlessly with real footage, or manipulate and enhance existing scenes by altering lighting, weather, or perspective. Variational Autoencoders (VAEs) can generate variations of the same scene or character, such as different expressions, poses, or costumes, and interpolate or extrapolate between frames to create smooth transitions or fill in missing details. Transformers can generate natural language descriptions or scripts based on keywords, genres, or plots, and produce audio or speech from text, emotions, or accents.
These generative AI techniques offer several advantages for filmmakers, including:
- Reducing production costs and time: Automating or simplifying tasks that currently require human expertise, equipment, or resources.
- Enhancing creativity and diversity: Enabling filmmakers to explore new ideas, styles, or genres, and create content that would otherwise be impossible or impractical to produce.
- Improving quality and realism: Generating content that is consistent, coherent, and detailed, and enhancing existing content with higher resolution, clarity, and fidelity.
Another significant application of generative AI in filmmaking is customizing or personalizing content for different audiences, platforms, or purposes by adapting or modifying existing movies or generating new content on demand. Generative AI models can tailor content to viewers’ preferences, interests, or needs, and meet the specifications, constraints, or goals of distributors.
For example, GANs can alter the appearance, style, or language of movies by changing colors, textures, filters, or translating subtitles, dubbing, or captions. VAEs can modify the content, plot, or genre by adding, removing, or replacing scenes, characters, or events, and changing the mood, tone, or theme. Transformers can generate summaries, reviews, or recommendations based on keywords, ratings, or feedback, and create interactive or immersive content, such as games, simulations, or virtual reality experiences.
To be clear, by 2025, generative AI will be capable of creating complete characters, sets, locations, camera angles, scripts, lighting, and voices, even learning stylistic nuances from famous directors. This technological leap will revolutionize filmmaking, much like human innovation has done for the past century.
What is the impact of generative AI technologies including generated video to Hollywood?
The technology is revolutionary for storytelling, creating market shifts at every turn. Currently, the average feature film costs about $120M for 2 hours, or $1M per minute. But what happens when that cost drops from $1M per minute to just $1 per minute?
Generative AI models like Sora, which use diffusion models to generate frame-by-frame photorealistic content, cost about $1 per minute of GPU time. This cost is expected to decrease by 50% every 12–18 months. Historically, when costs have dropped by such a magnitude, traditional methods and industries have been nearly obliterated. Consider the drastic reductions in computing and internet transport costs over the past 20–30 years — both have seen similar million-to-one decreases.
Take the cost of generating illustrations for a slide deck, for example. Just two years ago, hiring an artist would cost $1000 and take a week or two. Now, using a visual AI model, you can generate the same illustrations for a maximum of one penny (or perhaps $0.001 in GPU cost). This is a million-to-one reduction. If I were an illustrator, I’d realize this significantly impacts my livelihood, and I’d need to master these tools to remain relevant.
Similarly, consider a call center worker overseas, who averages about $1000 USD per month. Generative AI virtual assistants can now provide the same or better service for $10 per month per agent, with costs likely falling to $1 per month per agent within a year — a 1000X reduction. It’s evident that these jobs will become obsolete once AI takes over customer service tasks. For instance, Klarna famously laid off 700 call center workers and replaced them with AI in early 2024. This transformation is already underway.
Let me state clearly here (for those who are still wondering) that the impact to the film industry will be earth shattering. We know this because we can look back over the last 20 years and see many industries that were impacted profoundly (or decimated) by the internet.
Let me state clearly for those still wondering: the impact on the film industry will be earth-shattering. We can foresee this because we’ve witnessed similar transformations in various industries over the past 20 years due to the internet.
- The print media industry lost a significant portion of its readership and advertising revenue to online platforms like blogs, social media, and news aggregators.
- The music industry saw a decline in CD sales and a rise in digital piracy, alongside increased competition from streaming services such as Spotify, Apple Music, and YouTube.
- The retail industry experienced a shift in consumer behavior, with more people opting to shop online from e-commerce giants like Amazon, eBay, and Alibaba rather than visiting physical stores.
- The travel industry had to adapt to online intermediaries like Expedia, Booking.com, and Airbnb, which offered lower prices, more choices, and greater convenience to travelers.
- The education industry faced new challenges and opportunities from online learning platforms like Coursera, Udemy, and Khan Academy, which provided access to high-quality courses and materials, often for free or at a low cost.
- The photography industry saw the collapse of film in favor of digital, with prints being replaced by online sharing.
The change in the film industry will come quickly and be painful.
So where we will see impacts first?
- Eliminating the need for soundstages or perhaps any photography whatsoever. If I can generate any scene and any backdrop and any film location…why do I need a soundstage? Or cameras?
- Eliminating the need for actors. Digitally generated characters, that are photorealistic to humans, are already possible today. Why pay an actor $10M? Now you will say “because we love Tom Cruise” and we do, but you already have loved generated characters for decades from Mickey Mouse to Bugs Bunny to Shrek to Buzz Lightyear. Humans have already made the choice to enjoy and even love made up characters. Why not human looking ones? Guaranteed we will have generated characters who are stars you want to see again and again. Just like ones of the past that looked less human.
As a simple test, identify the AI generated faces below from the real human faces:
You cannot. Because these are generated with a GAN which continuously compares the generated image to a real human and provides feedback to make the generated one more believable until the computer can no longer identify which is real and not real.
Answer: These were all AI generated faces. And all can then be fully animated in realistic scenes. Today. More on that later.
Back to the impacts:
- Expanding creative possibilities for filmmakers, actors, writers, and anyone else by enabling the creation of realistic scenes and characters that would otherwise be too costly, risky, or impossible to produce.
- Reducing the need for expensive and time-consuming post-production processes, such as dubbing, editing, or CGI and visual effects, by generating high-quality content in real-time or with minimal human intervention. This could be a death blow to highly skilled CGI talent and studios, as AI can generate the desired effects using a diffusion model until perfected. Why spend $10M on a spaceship crashing into a planet when it can be generated for $1?
- Increasing diversity and representation of marginalized groups by allowing them to create their own stories and voices or modify existing ones without facing discrimination or bias.
- Challenging notions of authenticity, authorship, and consent by raising ethical and legal issues around the manipulation, ownership, and distribution of generated content, especially if it involves the likeness or identity of real people (hence the exclusion of real humans).
- Creating new opportunities for education, journalism, and social media by enabling the generation of engaging and informative content, but also posing risks of misinformation, deception, or propaganda.
- Empowering new voices and talents by lowering the barriers to entry and distribution for independent and aspiring creators who can use generative AI to produce and share their content online. This will enable anyone to make feature films for $120 and upload them to YouTube, resulting in visually stunning Hollywood-quality films. The quality of the stories will vary widely, with some creators gaining millions of views while others have none, and most content may be subpar. This new technology will democratize moviemaking, for better or worse. No soundstage, camera crew, post-production, or sound team needed. A person can generate a feature film in a few hours for virtually no money. Similar to how podcasting democratized broadcasting, where you no longer need to own a radio station to be heard — $200 is enough to get started. Most podcasts are mediocre, but occasionally a star like Mr. Beast emerges, making $64M a year. Feature film content will follow this trajectory.
Conclusion
I am a huge fan of Hollywood. From great storytelling to fabulous direction, camerawork, and editing, I love it all. Having been involved in filmmaking and active in AI for 25 years, I strive to offer a balanced perspective. However, history can be unpredictable. Or is it predictable?
By 2025, as generative AI models become widely accessible, the filmmaking process will be democratized. Anyone with a story and $120 will be able to create a film, with a product quality rivaling that of a $120M production. Given that distribution is already free (YouTube), the focus shifts to reducing the cost of making a world-class film. A great story (or premise) will still be essential, and the market will be inundated with both good and bad stories.
This influx of content — potentially thousands of movies daily — will drive down market value, making it difficult for Hollywood to distinguish its products. I predict that overall revenue in film and video production and distribution will plummet by 80% over the next decade, faster than the decline of newspaper revenue, dropping from $100B to $20B by 2035.
Despite this, we will empower millions of storytellers to share their narratives without the constraints of studios or financial limitations. We will witness AI-generated films being nominated for Oscars (or their equivalents) and even develop a fondness for human-like lead characters, despite knowing they aren’t real. Exciting times lie ahead.
