A New Narrative for "AI" in Entertainment

This article first appeared In Peter Csathy’s “The BrAIn” newsletter here.

"AI" is an enigma. Since it is not one single type of technology but rather an ambition, it is deliberately hard to classify.

And yet it is here, in various forms, encroaching on creative, administrative and operational work without yet a clear narrative for how it can be deployed in service of storytelling and the businesses we build around that.

Uncertainty breeds uncertainty. But this is a pivotal time for action.

We need a tangible approach to integrating "AI" into the entertainment industry driven by our own needs to solve real problems and leverage deep skills.

So here are three concrete ways AI can help optimize creativity and production  – three ways to form a practical narrative.

1: Story Creation

We face an undeniable challenge creating new visual stories today, because we operate a flawed story workflow. We build visual projects in text(!), so every collaborator has to interpret their own visual meaning from what’s written—and every contributor spends time correcting that vision instead of elaborating their story. We run scenarios on the surface because committing too deeply is time consuming, costly, and impractical.

Visual story creation processes should deliver instant feedback to creators as they build out each element of their story: opening, plot, character, mood. 

“AI” can in fact help here, by automating the process of text visualisation at any stage of development. Modern “AI” systems are fundamentally built on the structure of language and are increasingly capable of converting text to relevant images, running alternative scenarios based on human notes (‘prompts’). Iterations can be fast, maximising the creative options we can explore at the cheapest stage of production: development. Meanwhile all tools and their inputs and outputs can be tracked, enabling clear IP reporting, licensing and protection.

2: Production Planning

As we collaborate in development and prep we only see the true implications of our ideas for locations, scenes, shots, budgets, schedules, etc. late in the process. Often too late, when we are on set, with dozens or hundreds of people in tow. Or when we’ve already chosen our locations based on cast, crew, and financing. Prep feels wasted if it doesn’t end up on screen, so we focus on the most complex and costly of scenes, costumes, and props.

Production planning should integrate directly with the story creation process from the outset, so that every location option, shooting schedule, cast choices, and right down to individual shots are automatically defined or optioned based on the story’s iterations.

“AI” automations can be highly effective at stringing together distinct areas of knowhow to produce such combined suggestions, giving production teams a head start on the detailed work of bringing creative ideas to life, running alternative scenarios rapidly and in sufficient detail to be decisive.

3: Production Operations

Hundreds of processes must align to deliver the complexity of a film, TV, or commercial, covering creative, practical and administrative tasks and roles. Too often these are separated for most of the life of the production, uniting only fleetingly through manual and often distant connections in production meetings, emails, spreadsheets, contracts and accounts. This produces no unified overview of production, and no opportunity for continuous learning from each production to make the next one its best version of itself.

Production operations should be baked into the production process from the outset. 

Chain of title, financing and reporting should be monitored and maintained automatically by all activities undertaken to deliver the production. Creative freedom should be maximised within the clear boundaries of those constraints, and all teams should have access to the core production data, milestones and outputs they need to do their jobs.

In these three ways—for story creation, production planning, and production operations—we see how “AI” automation can make sense, supporting creatives, producers and studios to streamline existing processes and solve real problems. Automated visualisation enhances story development and collaboration, allowing for data-driven planning and real time adjustments. Responsive, transparent tracking and reporting can be built on top to help build collective knowledge for productions over time.

We need to shift our thinking from "AI" as a concept with endless possibilities, to a mechanism with specific capabilities. With only a little imagination we can harness “AI” automations for what they are good at and serve the needs of people. In the process we might just secure value and power in the things we create and open new methods for creating the best stories from all voices in every context. 

Until then, our reality means:

1. the Internet uncorked media distribution, moving power [audience + revenue] from separate legacy gatekeepers in production [studios] and distribution [theatres, TV, home video] to new integrated gatekeepers (streamers);

2. mobile accelerated this trend, introducing yet more gatekeepers in distribution (platform app stores) and audience aggregation (social media services), and fragmenting channels, formats, and audiences behind those gatekeepers and simultaneously disintermediating creators from audiences while mediating their value

3. costs grew from this new competition, investments were demanded for new business models, and COVID19 accelerated existing trends, making every production more expensive just as revenues dwindled, squeezing margins and giving the gatekeepers of consumer cash and attention (streamers, platform app stores, and social media services) ever greater collective leverage over creators and studios.

“AI” is actually perfect for the entertainment industry, because we are one of the few domains where humans are definitively in control of what is deemed high quality. You cannot outsource taste or opinion to a machine because all they do is to average what came before. So the outsourcing of certain work to automation can only exist within that wider edited, directed, and produced world run by human taste-makers.

The creative freedoms that automation can unlock are specifically valuable in entertainment, but not unique. Creative thinking is prized in countless other domains, so perhaps the entertainment industry can lead the way in how to adopt “AI” for human benefit, rather than displacement.