BroadwaysBroad said: "Could AI be used to create the projection designs and set designs and lighting designs? Would that put a lot of people out of work?"
Broadway LD here. I have some thoughts.
My job involves a LOT of work beyond just content creation. Attending meetings, giving feedback on script development, helping the director with world-building, making collaborative design decisions, interfacing with other departments...there's so much of my job that AI is not *remotely* close to being able to do.
Now, lighting and to a certain extent sound design are two of the most ephemeral design disciplines and probably the hardest to even imagine AI replacing. Let's talk about set design though. Could an AI model conceive of a set design? Sure - and you can ask generative AI models like Midjourney or Dall-E to do it for you right now. But they're not going to give you anything beyond renderings. A set designer's job is so much more than that. In addition to everything I do above, they have to think about the flow of a play, how is scenery going to move on and off, how are transitions going to work, how is the cast going to make entrances and exits. AI can't actually do any of this work - it's an extremely fancy Xerox machine. It predicts what is most likely to come next given a prompt. It can't problem solve like this yet, and it probably won't anytime soon.
Now...in terms of content creation...that's a different story. ELEMENTS of set designs involve generative work like painted backdrops, and projection designers and animators are constantly generating original artwork to use in shows. In that way, AI has already been helpful and has already made its way into shows. Has it replaced an animator's job? Probably not. Could it? Not anytime soon.
What about the crew? Is anyone at risk of losing their jobs on a show due to AI? I think there's a greater risk of losing jobs to general automation than to AI-specific models. Yet even there, we're running up against limits of the technology. Real-time tracking tech allows us to have moving lights follow performers with greater precision than ever before, but still not as good as a skilled stagehand on a follow spot (this is not a limitation of the brains behind the system, but rather the physical limitations of how motors work - they can't anticipate changes in direction and react like a human who know the show can). Similarly, there are scary things happening with how good sampling and electronic music sources can replicate orchestras, but that's not exactly new, and that's why AFM considers minimums such an existential piece of their contracts.
Obviously the dangers of generative AI are the ones we can't anticipate, and obviously the technology is in its infancy. But just given the very nature of how it works, how it is basically a very very very fancy text and image predictor, and given how the entire point of live theater is the LIVE part, I find it difficult to having a huge impact any time in the next several years. Definitely something we're all watching closely, though nowhere near as closely as our siblings in film/tv (where the risks and dangers to the art form are much, much greater).
Those are my under-caffeinated Friday morning thoughts.
