For many folks in M&E, the topic of generative AI is contentious.
The dust has hardly settled on Hollywood’s 2023 protests, and the AI protections in the resulting SAG-AFTRA agreement were dissatisfying for many people. Some studios have put an outright ban on using generative AI, for fear of facing costly lawsuits if the art it’s trained on is copyrighted.
Still, many studios are feeling the pressure to adopt AI for fear of being left behind, and others are already beginning to experiment with it.
It goes without saying that it’s here to stay, and your team may already be looking at how to use it for time and cost savings.
But before you overhaul your processes with AI, it’s important to level-set about what it can realistically be used for (hint: not replacing the creative process) and the infrastructure adjustments your team needs to make to take full advantage of it.
AI can let creatives do what they do best
First things first: Generative AI cannot, and will not, replace human creativity. We will always need people to manage and revise AI’s input and output. Otherwise, computers will be making content for other computers, and nobody wants that. In fact, a paper from Stanford and Rice University found that if you don’t feed AI enough original, human-made content, it breaks, and its outputs suffer drastically.
It doesn’t have the potential to replace a creative team, but it does have the potential to eliminate a lot of the tedium from creative work. For example, let’s say you’re creating art for a scene in which a thousand soldiers come storming over the horizon. You need to create a thousand realistic human faces.
Designing each of those thousand unique faces manually would take at least as many hours, and your team doesn’t have time or bandwidth for that, so that isn’t going to happen. Without generative AI, you may have to settle for a sea of copy-pasted faces that ruin the illusion and distract from the magic of the scene if your viewer looks too closely at them. With gen AI, you can make countless unique and realistic ones and automatically populate the models.
Now you end up with a scene that’s every bit as powerful as your team imagined it would be when you storyboarded it.
From generating preliminary mockups and rough storyboards, to renaming hundreds of files, to applying textures to thousands of files, AI can alleviate the tedium that plagues many creative workflows, letting artists focus on art — and making it easier for them to hold onto the creative spark that impactful content cannot be created without.
Aging tech stacks will crumble under the weight
AI is, and will continue to be, one of the most powerful and revolutionary tools in media and entertainment. Studios who are scrambling to make use of it are justified in doing so, but moving too quickly could create way worse problems for your organization down the road. You can’t add AI onto an outdated tech stack.
In my role at Perforce, I have worked with many M&E studios and found that teams, both small and large, have development pipelines straight from the 1990s and 2000s. In some cases, even large, established studios are having artists save files to their local machines, send them to each other via email, and have no secure way to share them with external collaborators. Many of these teams are already trying to implement modern technologies like real-time 3D engines, neural networks, and distributed render pipelines into their aging tech stack.
It’s no surprise they’re running into countless issues, from file conflicts to storage issues, to slowdowns when trying to share and collaborate on art assets.
Many studios’ tools and systems are already struggling to handle the immense data and iterations involved in creating a film or show, let alone all that’s involved in using the latest tech.
Many of their pipelines and workflows involve numerous manual processes, from the way they exchange files, track file versions, handle layers, and manage rendering. These inefficiencies compound into massive time sinks, in turn causing production delays.
And time equals money.
AI will introduce even more processes and exponentially more data into studios’ pipelines. For one, some studios will opt to train AI on their own IP to avoid risk of lawsuits, and doing this requires a tremendously large dataset. All these assets need to be catalogued and all the metadata tagged. On top of that, regardless of whether a studio trains its own AI or uses an out-of-the-box tool, they will have countless iterations of files to manage, each with its own trove of file metadata.
Spending additional dollars on solutions that will save them money over time directly conflicts with most studios’ actual yearly budget constraints. If they try to add AI to their outdated processes and legacy tech, though, they will risk their operations falling apart.
It will take more time to script, glue services together, and maintain those shaky pipelines than to look at their current pipeline and evaluate if it’s time to upgrade to a modern data management system—one that can support contributions from global and remote teams, easier reuse of IP, and automation within creative workflows.
Rapidly evolving tech calls for new, scalable systems
To make the best use of generative AI, studios need to first assess their existing infrastructure to determine if it could keep up with the massive influx of data. They need to ask themselves questions like:
• Where are we running into bottlenecks?
• Where are team members becoming stalled and wasting time?
• What are our current storage limitations?
• How can we iterate on content faster?
• Who will manage the large amounts of data generative AI will produce?
• How will data be stored, managed, and transferred?
They also need to consider if they have the expertise and resources to design and implement AI according to best practices.
While AI could be the key to accelerating your team’s creative workflows, scalable, secure, and highly configurable data management is the key to getting the most use out of it, and out of your own IP.
We’ve all seen how much an industry can change in a few short years.
Don’t skimp on establishing infrastructure that your studio can rely on for many years and advancements in technology to come. Version control is worth the investment.
===================
** By Ryan L’Italien, Gaming, M&E Evangelist, Perforce Software **