Generative AI's Use in Film and TV Pitch Decks

An exploration of the pros and cons of using AI generated visuals in pitching film and television projects.

One of the areas most affected by AI in the film and television industry is the use of generative imagery for pitch decks, and to a lesser extent, look books, mood boards, and reels. However, this adoption hasn't been devoid of controversy. It has prompted me to consider the relative effectiveness of using generated imagery in pitches and has sparked controversy regarding the legal, financial, and ethical issues it has raised. This essay delves into the advantages and drawbacks of employing AI-generated visuals in pitching film and television projects.

Legal Considerations: Is AI in a pitch deck breaking any laws?

Traditionally, pitches have relied on a patchwork of images sourced from various physical and digital sources, often comprising third-party copyrighted material that's stitched together to convey a project's essence—its world, characters, and aesthetics. The copyright holders are not aware that their work is being used. Pulling this material together is a real artform and, in the hands of a talented artist or designer, can significantly strengthen a pitch.

These presentations are typically shared behind closed doors with distributors, investors, lenders, talent, and their representatives. The public remains unaware of this foundational step in packaging and selling of a project. Given this precedent, employing AI in pitches seems akin to the traditional methods of pitch creation and presentation, thus arguably constituting an acceptable practice. Sort of a “if a tree falls in the woods and no one’s there to hear it” defense, but I think this is an apt description.

Financial Implications: Is AI "stealing from rights holders"?

Do AI companies or their users bear any responsibility towards the owners of the imagery used to train the AI tool? Using the classic tech start-up ethos, Generative AI firms quietly scoured the internet, set up shop, and awaited reactions. To date, these developers have defended their use of images published by artists by claiming these images fall under the fair use doctrine in the US. This doctrine promotes freedom of expression by permitting the unlicensed use of copyrighted material. US regulatory and policy initiatives are still struggling to catch up in establishing guidelines for the ethical use of AI technology.

Personally, I advocate for IP holders to have control over whether and how they're compensated for the use of their content. You’re starting to see blanket deals with major IP holders like the News Corps deal ($250M over 5 years) with Open AI where IP holders grant generative AI companies the right to scrape their content. Run the math on that and Open AI (and other generative AI companies) might be paying 100’s of billions for the rights to scrape all news, books, movies and music. Perhaps in the future, a blockchain solution could facilitate this process by tracking rights and automatically sharing the revenue resulting from the use of copyrighted IP. AI proponents maintain this would be massively cumbersome and the amounts being paid in most use cases would be infinitesimal. 

Ethical Considerations - Is AI replacing anyone’s job?: Before the advent of AI, our studio primarily relied on illustrators/concept artists for pitch deck imagery. It’s a relatively slow process with varying degrees of quality. In our studio’s particular case however, hand-drawn images often failed to align the aesthetics of the projects we were trying to portray. AI generated imagery got closer to the vision for the projects than hand rendered imagery so we adopted it in many of our pitches.

In the short term this will have an impact on artists' means of making a living. In the long  term, I believe that top-tier creatives will increasingly view AI as a valuable tool in crafting compelling pitch decks and give rise to a whole new class of artists that specialize in the use of AI tools to create generative imagery.

Efficacy - Do generated visuals enhance or detract from a pitch?

I vividly recall hearing reports from Cannes that everyone was walking around with AI generated imagery in their pitch decks. I remember thinking at the time - has generative AI already become a de facto prerequisite for a pitch deck? Or, was it already considered bad form to use AI in pitch decks?

First generation pitch decks with generative AI imagery have a quality that’s fairly easy to detect and the level of control was very hit or miss. They tended to have a gauzy, hyper idealized sameness and for that reason have slightly fallen out of favor at our studio as I write this. I already cringe a little when looking at some of our early decks with AI generated imagery.

The Process - Still clunky but improving

Rewriting prompts is cumbersome and time consuming. At times we’re forced to abandon images altogether because we cannot coax the tools to generate the image we’re looking for. Most images that made it into a deck required 20-30 (sometimes more) reruns before we landed on an acceptable image. 

It's been a learning curve with the tools evolving almost daily. The software we've primarily utilized, Midjourney, now facilitates style and character consistency across images—a pivotal advancement. This increased level of control and improvements in generating differentiated content that feels original, convinces me that AI will become standard in pitch decks.

In Summary:

Crafting high-quality, distinctive generative imagery for pitch decks is an art form. When handled by skilled designers, this imagery can significantly enhance a pitch. With increased control over the generation process, the potential for creating fresh, captivating visuals is immense. This evolution will likely lead studios to engage talented creatives specializing in AI-generated imagery. As a result, I would argue that the inclusion of generated imagery will become the new standard, ultimately elevating the strength and quality of pitch decks going forward.