Generative AI and the film production sector

Observations on the near future

As cofounder at Ambushed.tv, creating treatments for the top tier of advertising films, fiction, and series, I have had the opportunity to work with thousands of pitches and film briefs from all around the world, for various sectors (advertising, fiction, documentaries...). This gave me a deep insight into how both excited and afraid most film production companies are about Artificial intelligence. That doesn't make me a futurist, but here are some ideas about what we might see appear as trends as generative AI technologies double-up every month.

The democratization of filmmaking with AI tools

We’ll start with this assumption:  AI filmmaking tools, video generators, language models... will greatly democratize the access to filmmaking that was so far gatekept by production companies (due to the means it requires to organize high-quality footage and narrative). Some leaders in AI call it a decentralization effect, beneficial to creators (think Direct to Consumer, or Direct to Creators, cutting the middle man), but I believe this doesn’t take into account what truly happened in history in the past: I’ll write another article about the demise of most indie sound studios in the 2000s as platforms like Spotify and Deezer and the followers were advertising this form of decentralization, but we know that it ended up in a centralization of their powers.

Concretely, toys-tools as I call them, Runway, Stable Diffusion’s latest mini-tools like Turbo, Midjourney, and others to come (like Canva for Web-Designers, as opposed to Photoshop, Adobe suites, complex website builders...) will allow “hobbyists” to create great narratives and stories, almost super-realistic and Hollywood-like or sometimes quite trippy content that will compete with what production companies have been able to create so far. The trend then, could be a tsunami of user-generated content, for the most of average quality at first, and getting better, focusing on specific style niches, tribes, bubbles... The future will be so exciting for small content creators willing to share their local stories with the rest of the world.

The Hypothesis

VFX will become cheaper but at a price

When talking about the inherent risks and opportunities generative AI creates for VFX professional editors and makers, there seem to be 2 doctrines here: some think it might enhance the ability of a small percentage to create better content and make the rest, the “middle-class”, redundant, in a vision where AI permeates throughout the content creation world. Some others think it will never be able to replace the human hand and will add on top of the traditional sector to create a new genre, a new form of visual communication. I believe that both arguments have some truth in them and we’ll see a right balance of both come to play.

The lack of macro control over the final result that is generated with AI (as opposed to crafting VFXs or 3D worlds by hand) will allow people to deliver content that is more creative, more imaginative, and more trippy, but never exactly what they might have in mind when they started their project (I’ll treat this subject, the ratio control/imagination expansion that occurs with the creative chaos that AI delivers).

Post-production brings the next unique advantage

Post-production is already more and more important compared to the past in the way we create films. This transformation already happened with large incumbents that had the means to finance the next tech in filmmaking and post-production systems. But with AI, access to these tools gets exponentially cheaper and some processes are almost fully-automated. The access to large amounts of computation power and data will widen the gap between middle-sized production companies and the largest consortiums. This will probably also widen the gap between cheap and cheerful VFX (that can have its charms) and top-of-the-edge VFX (also assisted by AI, but controlled by humans in the execution) that will require heavy computation power and access to very complex machines to “guide” the AI and make it controllable and awesome.

In a nutshell, expect more animations, motions, illustrations, and blends of them. We see this trend that follows a strategy that Hermes put in place a long time ago, luxury fashion brands following now (Audemars Piguet...) - a mix of motion and live capture might be a real big trend.

Genres and sub-genres will have their own platforms and specific content

Visual narratives will continue to fragment into genres and subgenres: the future is all about ultra personalization and very tiny tribes, therefore content that also engages with how specific each tribe communicates, likes, and thinks.

IP is (not?) dead

A multiplication of reproductions, concept thievery, extrapolation, and other mutations will occur at a pace never seen yet, giving even more hardships to original content creators. I’ve personally witnessed plenty of prototypes or pre-creations shared briefly amongst communities and immediately repurposed by other creators that published first and therefore, captured all the attention. It seems that the fastest creators will thrive in this new paradigm. We also have to accept this effect of recuperation and harness it: create more, focus less on protecting your creations seems to be the coming motto.

Open source tools will be the main tools for Film production companies willing to keep the cool edge

Open-source tools are the base for both toy tools and creators that can harness their extremely complex and complete adaptability. Runway, Midjourney, and other tools are already based on the work of a massive community of open-source programmers and creators that exist in a parallel world, in shadowy Discord servers. Open-source tools like Comfy.UI, Automatic1111, and the new tech releases that come with them every week will allow a minority of creators to create mind-blowing films and stay on top of the narrative.

Meta suggestion: New universes and cosmogonies will appear

The ability to imagine things further than any time before by crushing loads of data and reshuffling the underlying systems we had in mind is going to be the next move: people will want to see complex, deep, ultra-imaginative content that will blow their minds.

Creators will have to think even more as a collective

A focus on the capacity to craft unique and refreshing narratives with the aid of AI Generative tools, as a collective of humans truly harnessing the synergies, enhancing our ability to be more human in the face of the machine. This concept might sound somewhat ambitious and vague, yet it is a key topic of my research with various groups of creatives and production companies.

The gap to be filled between great taste, real narratives, human-based creativity, and originality (Imagination, creative chaos) will be the most amazing playground for film production companies that stay nimble and understand the tools, allowing them to grow into this new world.

At synest.ai, we are creating a full-stack advisory and transformative service for the AI revolution