AI as a Tool
Assisting with Pre-Production
AI is being used to streamline and help manage pre-production pipelines:
Plan Schedules
Location Scouting and Planning
Automate schedules and shooting plans
Timetables and availability of actors, props and locations
This is all to support and increase efficiency and ultimately save money and additional costs that could be incurred with a longer pre-production process.
Scriptwriting
We are already seeing the impact that AI is having on scripts and the backlash that this can bring but principally AI is being seen by studios as a way to reduce costs and streamline processes thus maximise profit. Using machine learning scripts can be refined by analysing data and impact of previous test audience reactions to script pacing and cohesion whilst also keeping continuity.
Actor Selection and Audition
Actors can be selected and audition using AI tools that compare facial features, emotive responses and specified criteria against requirements and review these against databases of actors and availability. This when combined with deepfakes and meta human technologies allow studios to digitally emulate actors faces over digital / body doubles before committing to any casting. This could also allow for entirely digital actor doubles to be used with AI to copy vocal ranges, expressions and acting neuances for characters however this does bring into question ownership of self and image alongside employment concerns.
This is more cost effective in many ways as less time comparing actors and working around schedules may be needed however brings up multiple ethical concerns.
Predictions and Reactions
Film and TV production studios ultimately want to produce content that people want to watch and at least break even on their budgets as such being able to analyse what will and won’t work based on trends, other content reactions and general feedback is critical and this is where AI is being used increasingly. Analysing film scripts, actors reactions and general trends and interest using machine learning whilst not a perfect science can curb extra costs and try and make content people want to see. Engines such as ScriptBook, Merlin and Cinelytic are all used by major film studios to predict how well a film will do sometimes before even a single frame is shot.
Editing
Editors can use AI to cut, edit and otherwise process their videos along a timeline using AI to identify key moments, remove errors, upscale footage, remove background noise and so much more streamlining workflows that may take hours and days. We have seen AI in editing packages for years from facial recognition to auto tracking for VFX but the ability to dynamically change the context of a scene and add / remove elements has only gotten better requiring less and less manual work.
Trailers can be cut together on story beats to maximise audience impact and draw alongside using facial and scenic analysis to collate all scenes with set actors and scenes together to make editing easier alongside auto generating sub-titles, dubbing and credits.
Audio Production
How a movie or show sounds is as important as the visuals on screen and AI can be used to compose music, translate vocals, and otherwise match audio cues to visual actions to streamline editing and generate music without the need for composers or studio time with its own pitfalls. AI can already emulate other people’s speech patterns and tone and with dubbing tools can even adjust mouth movements to change dialogue to make it seem as if an actor has said something they did not.
Making Movies
We have seen a rise in AI-generated video content whereby entire scenes have been created entirely via prompts and whilst the results of these have been mixed as the technology improves the sheer amount of generated content that can be produced on a small budget in a short amount of time is something to be aware of. In 2016 an AI called ‘Benjamin’ was created by a couple of filmmakers working as creative technologists at Google whereby they created a film script within 48 Hours (Film: Sunspring) and since then they have continued to refine their engine to make further films. Since then generative moving image AI has vastly improved and whilst somewhat surreal at times we have already seen its results enter mainstream TV and film such as the Intro sequence for Secret Invasion.