AI Goes to Hollywood

New technology, of all kinds, is notoriously difficult to “put back in the box” once it’s out. Early-stage technologies tend to expand rapidly, often outpacing society’s ability to adapt. As a result, initial responses to these rapid advancements are typically reactive and focused on potential negative impacts.

The 2023 Hollywood strikes exemplified this phenomenon. While the labour dispute was rooted in the effects of streaming on the industry’s economic model, the perceived threat of Artificial Intelligence (AI) to filmmaking became an increasingly prominent issue.

But were the screenwriters and actors right to be concerned? The answer, as is often the case with emerging technologies, is complex. AI will undoubtedly change Hollywood, just as it will transform all industries. Let’s take a deep dive into the film industry to explore the potential effects of AI.

Screenwriting

From Campfires to Computers: The Evolution of Storytelling

To fully appreciate the role AI might play in the future of filmmaking, we need to step back and consider the broader context of storytelling itself. Storytelling is perhaps humanity’s oldest art form, a fundamental aspect of our species that predates written language, agriculture, and even permanent settlements.

The evolution of storytelling mirrors the march of technological progress:

1. Oral tradition: Stories were fluid, adaptive, passed down through generations by word of mouth. Each telling was a unique performance, shaped by the teller and the audience.
2. Writing (c. 3200 BCE): For the first time, stories could be fixed, preserved exactly as their authors intended. This permanence allowed for more complex narratives, but limited access to the literate elite.
3. Printing press (15th century): Books could be mass-produced, making them affordable and accessible to a much wider audience. This laid the groundwork for the novel.
4. 20th century media: Radio, television, and cinema brought new dimensions to storytelling, combining visual and auditory elements.
5. Internet age: New forms like interactive narratives and transmedia storytelling emerged. Content creation was democratised to an unprecedented degree.

Each of these technological advancements expanded the reach and accessibility of storytelling. They provided new tools for creators and new experiences for audiences. But crucially, they didn’t replace human creativity – they enhanced it.

AI represents the next step in this evolution. But unlike previous technological advancements which primarily affected distribution and consumption, AI has the potential to both change distribution and impact the creation process equally. This is what makes it both exciting and terrifying to many in the industry.

The Limit of AI’s Imagination

AI’s ability to assist in creative processes is undeniable, yet it faces inherent limitations in generating truly original content. At their core, AI models—be they Large Language Models or image generators—are sophisticated pattern recognition systems, bound by the confines of their training data.

Consider an AI trained exclusively on black-and-white films from the early 20th century. This AI, despite its processing power, couldn’t conceive of colour cinema, let alone today’s immersive 3D experiences. This isn’t a technical shortcoming, but a conceptual one—the AI can only operate within the boundaries of its known world; the data that it has been trained on.

This limitation becomes critical when we examine AI’s role in creative fields. While AI can produce seemingly novel content, this novelty is merely a recombination or extrapolation of existing data. The truly groundbreaking ideas—those that redefine genres and push storytelling boundaries—remain out of AI’s reach, a direct consequence of its fundamental architecture.

As a result, the ability to create something original, films that transcend existing paradigms and construct entire cinematic universes from scratch, remains a uniquely human ability.

Avatar: Pushing the Boundaries of Human Creativity

James Cameron’s ‘Avatar’ is a testament to the power of that originality. Released in 2009, it became the highest-grossing film of all time, a feat nearly matched by its 2022 sequel.

What makes Avatar’s success particularly striking is that it wasn’t an adaptation, remake, or comic-book franchise instalment; it was an entirely original film. Yes, critics pointed out echoes of familiar narratives like ‘Pocahontas’ or ‘Dances with Wolves’, but this missed the forest for the trees. The true innovation of Avatar lies not in its plot, but in its execution – the immersive world-building, the groundbreaking visual effects, and the overall cinematic experience that Cameron crafted.

Cameron didn’t just sit down and write a script. He spent years developing the world of Pandora, drawing inspiration from Earth’s deep-sea ecosystems, which he explored in his documentary work. He collaborated with linguists to create the Na’vi language, with artists to design the flora and fauna of Pandora, and with engineers to develop the technology needed to bring his vision to life. He pushed the boundaries of what was technologically possible in filmmaking, pioneering new motion capture techniques and 3D cinematography.

This level of holistic creativity – combining lived experience with storytelling and technological innovation – is something that current AI, bound by its training data, simply cannot replicate.

Human Connection

That ability is not limited to film; the intricacies of the best long-form television series demonstrate that the best storytelling is predicated on building long-term human and emotional connections with audiences. Shows like ‘Breaking Bad’ and ‘The Wire’ weave complex narratives that span dozens of hours, balancing episodic stories with overarching season arcs and series-long character developments.

To replicate just this style of storytelling, an AI would not only require an enormous context window to manage entire season and series storylines – a feat that’s far beyond current capabilities – it would need to understand and replicate the subtle art of pacing, the careful revelation of information, and the delicate balance of multiple character arcs that make these series so compelling.

AI models, despite their complexity, lack the deep understanding of the human condition required to create these types of work. On the surface, ‘The Wire’ is a crime drama about the Baltimore drug scene. But over five seasons, it evolves into a profound exploration of institutional dysfunction, touching on education, politics, media, and the death of the American working class. The show’s creator, David Simon, drew on his experience as a crime reporter to create a narrative that was both entertaining and deeply insightful about American urban life.

Storytelling in this manner requires not just creativity, but also real-world knowledge, emotional intelligence, and a nuanced understanding of human society – attributes that AI, which can only work with the tangible data it’s been trained on, struggles to replicate.

Focusing on the best of human filmmaking and comparing it to the output of AIs is possibly a little disingenuous. These films and TV series don’t represent the mean, and certainly don’t represent the entire spectrum. By their very nature, they are the ‘standouts’.

There is a significant market for what might be termed ‘filler’ or ‘blockbuster bargain bin’ content; the lower-end, Hallmark channel comedies and rom-coms that follow well-established formulas. It is here that AI will likely have the largest impact on screenwriting. The ability for AI models to produce scripts that follow a well-trodden path already exists and AI will likely lead to a glut of these straight-to-streaming action movies or holiday rom-coms that follow well-established formulas.

Performance

This turns our attention to the performance aspect of filmmaking – the work of actors and directors in bringing a script to life. It is here that AI’s role becomes both more prominent and more contentious.

Quality Acting

The idea of AI or Computer-Generated Imagery (CGI) replacing human actors is not new territory.

Indeed, we’ve already seen impressive use of digital de-aging technology in films like ‘The Irishman’ and the creation of digital versions of deceased actors in ‘Rogue One’. However, it’s crucial to understand that these digital creations are ultimately based on human performances. In ‘Rogue One’, the digital recreation of Peter Cushing’s Grand Moff Tarkin was based on a motion-capture performance by actor Guy Henry. Computers didn’t create the performance – they merely helped translate a human performance into a digital model.

AI could dramatically enhance these capabilities, making it easier and more cost-effective to alter actors’ appearances or even create fully digital characters, but while it can replicate, manipulate, and, most recently, completely animate, it will continue to struggle with the kind of spontaneous creativity and emotional nuance that defines great acting. The chemistry of great casting matchups–Bryan Cranston and Aaron Paul, Brad Pitt and Angelina Jolie, or Simon Pegg and Nick Frost–is hard enough for humans to emulate, let alone for an AI.

We’re focusing on high-end cinema though; the best performances from the best actors. As with screenwriting, the harder effects of AI are most likely to be felt in the mid to lower end of the profession. With AIs able to do a passable job on generating performances at relatively low cost, many of these types of studios will likely turn to AI for a significant portion of their performance work, even if it is not the performances of the main characters.

The main cast are not the only actors to consider though; extras-heavy scenes, already often created in post-production with VFX, such as street or crowd scenes which would once have required a location shoot and dozens of extras and props, could now be generated in a snap.

With the ease of production of these types of films, we may see lower-end studios opting to put out tens of films a year produced in this way, or even see new types of hyper-personalised films emerge: A single base film could, via AI, become 1,000 different variations to appeal to niche markets. One single script and ‘shoot’ with human actors could result in many different films which appeal to various niches, with the base performances modified to feature characters with different hair colours, environments or even ethnicities.

Voice

One area where AI might have a more transformative impact is in voice acting and dubbing. AI voice synthesis is advancing rapidly, and we might soon see AI systems capable of translating and dubbing films into multiple languages with the original actor’s voice while syncing the lip movements of the actors. This could make foreign language films more accessible to global audiences and potentially transform the economics of international film distribution.

Yet even here, the human element remains crucial. Voice acting isn’t just about producing the right sounds – it’s about bringing characters to life, conveying emotion and subtext through vocal performance. While AI might be able to handle more straightforward dubbing tasks, high-quality voice work is likely to remain the domain of human actors for the foreseeable future.

Solo Spielbergs

One of the most exciting prospects of AI in filmmaking is the potential democratisation of high-quality content creation. AI tools could empower solo creators or small teams to produce today’s-Hollywood-calibre films from their bedrooms, dramatically lowering the barriers to entry for aspiring filmmakers. Very soon, individuals with creative vision but limited resources could compete with established studios in terms of production quality.

Imagine a world where a teenager with a compelling story could use AI tools to help them generate a screenplay, create stunning visual effects, and even produce realistic voice acting for multiple characters. MKBHD talks about needing to be an octopus as a solo creator - AI will give those creators access to robotic arms allowing them to create work that would otherwise be impossible by themselves. This level of democratisation could lead to an explosion of diverse voices and stories in the film industry.

Post-Production

Behind the scenes, the story becomes more complicated. While the past decades have seen huge growth in the world of VFX, it is apparent that AI will soon be an important tool within the arsenal of art directors.

Whilst current AI video production has issues with consistency, morphing and physics, the field is moving extremely quickly.

At the very least, it is now at a point where it can convincingly produce ‘B-roll’ footage which can then be edited into something that works. This B-roll or background footage may have been shot by the ‘second unit’ or created in VFX, which will have a labour impact on those teams. How quickly it gets to the point where the more fine-grained control that is required for professional productions is available, however, is difficult to say.

As AI tools become more prevalent in filmmaking, we’re likely to see a significant evolution in what constitutes spectacle on screen. When every production has access to stunning visual effects and the ability to create fantastical worlds with relative ease, the bar for what impresses audiences will inevitably be raised.

The audience should be the beneficiary, with a potential arms race of visual innovation, with big-budget productions pushing the boundaries of what’s visually possible to differentiate themselves or the emergence of new forms of visual storytelling that take full advantage of AI’s capabilities.

With more creative freedom for filmmakers, as AI tools reduce the cost and complexity of realising their visions on screen, we may see a new wave of visually ambitious independent films, as creators who previously couldn’t afford extensive visual effects gain access to powerful AI tools.

On the other hand, as audiences become accustomed to high levels of visual spectacle across all types of content, we might see a renewed emphasis on storytelling and high-quality, real acting. When every film can offer stunning visuals, the ones that stand out will be those that offer something unique in terms of narrative, character, or thematic depth. Just as the rise of processed foods led to a renewed appreciation for artisanal, handcrafted products, the proliferation of AI in storytelling could lead to a counter-movement celebrating distinctly human creativity.

This could manifest in various ways; a continued resurgence of practical effects in film, with audiences appreciating the tangibility and imperfections of real-world stunts and sets, similar to the approach of both Christopher Nolan or Tom Cruise; a renewed interest in single-take shots or long, unbroken scenes that showcase actors’ raw performances and a renaissance in documentary filmmaking, with its focus on real human stories and experiences.

Distribution and Consumption

All of these shifts in the creative process could also necessitate entirely new distribution models.

With solo creators able to publish full films, we might see the emergence of aggregator platforms that combine elements of Netflix, YouTube, and social media, allowing creators to publish their works directly to viewers and leveraging both user feedback and AI-driven recommendation engines to match content with interested audiences.

The democratisation of film creation also presents challenges. As the barriers to entry lower, we’re likely to see a flood of content. This could make it even harder for truly original and compelling stories to break through the noise. The challenge of discovery – already a significant issue in the streaming era – could become even more pronounced. Unlike TikTok where the investment of time a creator asks for is 30 seconds to a minute, films require a greater willingness from the audience to engage with an unknown quantity.

Moreover, this democratisation could significantly impact the economics of the film industry. The traditional studio system, with its high barriers to entry and control over distribution, could face unprecedented challenges. We might see a shift towards a more decentralised model, with a long tail of independent creators alongside the big studios.

New ways of consumption that have deeper integrations with AIs that are, via smartwatches and other data sources, able to understand a viewer’s emotional state and environment, might allow platforms to either offer content that accords with that, or even to specifically generate content that works for you.

Such a platform could use AI to analyse viewer preferences and behaviours at a granular level to suggest the perfect content for each viewing session. This could create new opportunities for niche content to find its audience.

More interactive content, whether directly, in the same vein as Netflix’s Black Mirror: Bandersnatch, or indirectly, with a film adjusting its pacing, tone, or even plot points based on real-time analysis of viewer engagement, could allow users to enjoy a new type of media that sits somewhere between a traditional film or a game.

We could also see the emergence of more personalised narratives, where AI tailors the story to individual viewers’ preferences, even going so far as having films that are adaptable to the individual viewer on the fly; a hero that looks like you, in the world that you know.

I doubt that the communal cinema experience is going anywhere anytime soon though and expect that there will be an increase in more immersive, IMAX-style, blockbuster experiences which aim to offer a highly differentiated product to audiences.

The Economics of AI in Hollywood

The integration of AI into filmmaking will inevitably reshape the industry’s economics. This shift is likely to be as significant as the rise of streaming platforms, if not more so.

One of the most immediate impacts will be on production costs. AI has the potential to streamline nearly all aspects of production, especially in post-production.

The accessibility of complex visuals to all teams could also be a boon to mid-tier studios though. This could make visually ambitious films more feasible at lower budget levels, potentially revitalising the mid-budget sci-fi and fantasy genres.

The Hollywood strikes weren’t wrong; there will be an effect on the labour force, from scriptwriting to visual effects and everyone in between. It’s not all bad news though; while some roles will be displaced by AI, new positions focused on AI implementation and oversight will emerge. We might see the rise of ‘AI wranglers’ on film sets, experts who can interface between traditional filmmaking roles and AI systems.

The potential for AI-driven personalised content could create new revenue streams but might also fragment audiences. If viewers can get hyper-personalised content tailored to their specific tastes, will they still be willing to pay premium prices for big-budget, one-size-fits-all blockbusters?

As AI becomes capable of generating content based on existing intellectual property, we might see changes in how IP is valued and protected. The ability of AI models specifically trained on a franchise to create new stories in established universes could make these already valuable franchises even more lucrative. The flipside is that, as we have seen with Disney’s mixed success in their huge growth of the Star Wars universe, fans sometimes actually prefer it when they don’t feel that their favourite stories are turned into derivative and low-effort output.

Moreover, as AI-generated content becomes more prevalent, we might see the value of original, human-created stories increase. Just as the rise of streaming led to bidding wars for top showrunners and content creators, we might see studios placing a premium on distinctly human creativity as a way to stand out in an AI-saturated market.

The international market, already a crucial consideration for major studios, could become even more significant. AI’s potential to easily dub and localise content could make international distribution more seamless, potentially leading to more globally oriented production strategies.

AI will present challenges to Hollywood and the filmmaking industry, particularly in terms of job displacement and market disruption, but it’s also going to create new opportunities. The studios and creators who thrive in this new landscape will be those who can effectively harness AI’s capabilities while maintaining the human creativity and vision that ultimately drives the art of cinema.>