Industry AI Guidelines Don't Have to be Arbitrary
"I don't want to be just one thing. I can't be. I want to be brave, and I want to be selfless, intelligent, honest and kind. Well, I'm still working on kind." – Four, "Divergent," Summit Entertainment, 2014.
We're moving hellbent into a new, different industry with everyone intent on using generative AI to its fullest. But we're not certainly what it's fullest is and whether it's all good. Just not certain if some of it we'll be able to fix in post but guess we'll soon find out. Right?
Recently, we were chatting (okay, texting) with a friend who mentioned an article we had read about things that were once great but are now lost.
One comment was a claim that shows/movies were far better in the old days compared to all of the new options. We said it might be the overflow of content today. According to Statista, only about 200 scripted projects were released in 2013 but last year, there were nearly 600.
He proceeded to enlighten us:
Hollywood shows/movies weren't inclusive back then
They didn't portray any reality
War movies were stupid propaganda
Shows/movies were heavily censored
Content was strongly controlled by a few studios
A lot of crap was produced to please studio/network bosses
Shows/movies back in the day sucked
We were hoping he would tell us what he really thought.
Okay ... maybe.
However, 10 years ago, it was almost a badge of professionalism to work in the writers' rooms or on set 10 hours a day, including Fathers day.
Actors, background actors/extras, stunt personnel struggled inside heavy costumes and sweaty make–up.
Entire crews worked in freezing cold or sweltering locations to ensure "authenticity" and exotic locations weren't exotic.
Postproduction folks constantly pulled all–nighters to meet abbreviated deadlines.
You could sleep, heal (bones and egos), recover when the project was delivered.
The budget (time/money) and audience ratings/response were what really mattered.
Today, new hardware/software technologies are making the work better, easier, more creative, more productive for everyone.
One of the best, we feel, was LED virtual production.
They've virtually eliminated the need to move everything/everyone for location shoots.
The technology made it possible to deliver "real" environments that only existed in the minds of the creatives.
Everyone saw what the camera saw.
"We'll fix it in post," was virtually eliminated from the discussion.
Developed and popularized by ILM (Industrial Light & Magic) for the production of Star Wars and the TV series, The Mandalorian, the wall puts the actors into the realistic virtual environment, so you don't have to use your imagination ... it's there.
Stage of Choice – The LED virtual production technology has been a dramatic leap forward for show/movie production. Directors, cinematographers and actors don't have to imagine their surroundings as with previous green screen productions because they're right there. They can visualize what the project will look like which saves time, money and frustration.
Of course, it relies heavily on Unreal Engine's 3D game engine, Nvidia GPUs (graphic processor units) and powerful computer systems to deliver the environment; but that's just techie stuff.
The virtual production wall has almost eliminated the need for the construction of large, expensive and complex sets.
The technology resulted in the loss of some construction jobs but most of the crew gained new skills and adapted to delivering the new set environment.
The LED virtual technology is such an improvement in content production that all of the studios being built in the Americas and around the globe include them.
With new content production facilities being built in the Americas and around the globe, one of the first things content creators/producers and studio owners want to know is if they offer the virtual production environment.
The technology provides a more realistic and immersive environment that enables real–time set changes, streamlined filming, on set/in camera adjustments and faster/less costly postproduction work.
The technology has become so reliable (and economic) that one of our local stations introduced the new environment into their newscast and production workflow.
Localized – Virtual production technology has become more reliable, easier to use and more economic. Even news/production stations across the country like KPIX in San Francisco have begun using it to differentiate themselves in their market and attract/retain their audiences.
The implementation of virtual/augmented environment is an important addition for one of our local news operation.
After all, they have to compete against social media and streaming content for audience attention.
All of the facilities are being used for the same reason ... put seats in theater seats and capture and retain home viewers.
Valued at nearly $3B last year, industry analysts project that the virtual wall and production market will reach more than $5.1B by 2027.
The technology has not only improved the entertainment quality of films and shows for viewers, it has also had a positive impact on the lives and performance of people who create shows/movies.
AI could have a similar positive effect on the creation of shows/films, but it has to be released/used in a cautious/controlled manner.
Before the end of last year, Sam Altman was fired and then reinstated as CEO of OpenAI and a new board of directors was voted in ... hey, stuff happens.
Helping Hand – If it feels like everyone is developing a powerful (profitable) AI tool but you, you're not wrong. OpenAI and many of the emerging firms focus on developing and releasing the products as quickly as possible without a lot of testing to find issues and problems before they are widely used. Some of the products will be something that keeps on giving in all the wrong ways.
His approach was similar to that of a lot of software developers ... develop it, release it, break it and we'll fix it when folks find a problem.
That worked fine when you developed/delivered a video game or work processing package.
It doesn't work with AI – which has been quietly automating human tasks for years – because the foundation of the technology is that it simulates the human intelligence process.
Last we checked, the basic goal of the human intelligent process is to ... survive.
We realize the AI tools are developed to do special things better and more reliably or help people do their work more quickly and accurately but ...
We know, cheap shot.
But the EU has already spent considerable time looking at the technology from almost every angle and have developed a detailed set of industry/application guidelines – https://tinyurl.com/3jdtm6nd.
They aren't as encompassing as people will want but ... it's a start.
In addition, the high–level guidelines can provide the foundation for the content industry.
The best solution for the film/video story industry would be for international professional standard organizations like HPA (Hollywood Professional Association), SMPTE (Society of Motion Picture and Television Engineers) and similar organizations around the globe to spell out a specific set of guidelines for the worldwide entertainment industry.
They won't be perfect, and they won't cover all applications, but they would provide the industry with a general set of guidelines organizations could adapt/adopt to meet specific needs.
Enhanced Postproduction – Major postproduction tool makers have cautiously implemented AI technology into their products to make even better films/shows. The new tools make it easier and faster for editors and their teams to deliver even better video stories while reducing time/financial budgets.
Even without these guidelines, a number of industry leaders including Adobe, Avid and Blackmagic/DaVinci have already introduced AI–enabled postproduction tools that professionals are using to edit, color correct/grade and add/refine visual effects.
The AI–enabled tools haven't replaced members of the post team, but they have enabled folks to deliver outstanding video, good professionals to produce great content and junior staff members to do good work at the outset.
In addition, even before remote work became "a thing,", project postproduction was being done globally.
Using a secure cloud and common tools, it's very common for different professionals in Atlanta, Berlin, India, New Zealand and Burbank to work on a project at the same time, passing segments from time zone to time zone and specialty to specialty.
As a result, the global team can stay focused and still produce a superior video product quickly and economically.
With virtual production studios and post teams located around the world, streamers like Netflix, Amazon, Apple and Disney can acquire great video stories "over there" and make them available here or ... anywhere.
Global Production – Studios/producers are able to work with post professionals around the world passing work from one team to another while still maintaining quality.
Historically, localization of films and shows has been a major hurdle for video projects to be enjoyed by people in Germany, Mexico, India, Dubai and New York City.
Translating the audio and adding the written translation in the lower one inch of the screen was usually the fastest and most economic approach.
However, accurate translation has been difficult and time–consuming.
Dubbing (having the actor "speak" the audio in another language requires precise mouth/lip manipulation with a voice actor saying the dialogue so it looks/sound realistic) has been very time–consuming and expensive.
AI–enabled localization tools have dramatically improved the quality of the dubbed/subtitled projects while reducing the time and cost required to perform the task to make the total film/show interesting and immersive for the viewer regardless of his/her native language.
Localization – AI tools have become indispensable in localizing films/shows in all of the major languages/dialects. The tools make the work faster and more reliable than yesterday's manual processes.
While there are more than 7,000 languages spoken around the world, Allan McLennan, president of 2G Digital Post noted that more than half speak the top 20 languages.
"With apps like our AI translation/localization tools, we're able to rapidly, economically and most important accurately make a studio/streamer project produced in one country available for release in other countries so the actors speak, sound and express themselves naturally in the audience's native tongue. he explained.
Native/Subs – Many viewers find a film or show presented in the native language, combined with easy–to–follow subtitles more satisfying and more enjoyable. We find it an excellent way to experience and follow the video and audio story.
"Our offices in key locations around the world have processed vast amounts of text, acronyms, abbreviations, voice and dialect material so our localization tools can "learn" to speak and present the content accurately, believably," McLennan continued.
"The prior labor–intensive work was expensive and often prone to poor or inaccurate interpretation for subtitles. It was also extremely difficult to provide consistent lip–sync for the dubbing while the audio was limited to the specific voice actor," he added.
"No one was happy with the results," McLennan concluded, "especially the audience."
But once the industry has a basic foundation of AI guidelines that has been set by the industry standards groups – SMPTE, HPA, etc. – then the film/show industries' creative and production folks – WGA, SAG/AFTRA, other guilds/unions – can use the industry foundation as a logical and professional base to define the specific AI guidelines that meet everyone's goals.
Will the final guidelines/guardrails meet everyone's needs/expectations?
Of course not. Some jobs will be displaced while new openings/opportunities emerge.
But it will sure beat 100 days without pay; and more importantly, stepping off the edge and hoping there's a safety net there for you.
If there are a clear, concise set of global and industry application and do/don't guidelines for AI, at least you know you're not in the wrong place ... at the wrong time.
Just remember what Four said in Divergence, "Fear does something strange to people like Al. But not you. Fear doesn't shut you down, it wakes you up."
Maybe with the right guidelines that everyone follows, we'll be able to develop and produce interesting/inviting shows/movies, get paid properly for them and be able to have sometime to relax, enjoy them.
Andy Marken – firstname.lastname@example.org – is an author of more than 800 articles on management, marketing, communications, industry trends in media & entertainment, consumer electronics, software, and applications. An internationally recognized marketing/communications consultant with a broad range of technical and industry expertise especially in storage, storage management and film/video production fields; he has an extended range of relationships with business, industry trade press, online media, and industry analysts/consultants.