AI in Show Design

[cybrunette]

It's only a matter of time before someone or a group of people start feeding AI enough information to reference for these AI engines to generate shows & scores that are clean enough to blend in with even the most competitive marching bands & corps. I also wouldn't find it surprising if someone decided to start a Discord server just for this purpose. 

Is this something to worry about or is this something to look forward to? Would we eventually see a plethora of new rules in response to AI in competitive circuits? Maybe even having reaching the point of bands having to produce & provide a paper trails of who or what designed their show? 

[supersound]

I honestly genuinely do not see this becoming an issue.

[allthingschaotic]

Jun 6, 2023 15:46:04 GMT -6 cybrunette said:

It's only a matter of time before someone or a group of people start feeding AI enough information to reference for these AI engines to generate shows & scores that are clean enough to blend in with even the most competitive marching bands & corps. I also wouldn't find it surprising if someone decided to start a Discord server just for this purpose. 

I've already seen some people ask ChatGPT for show ideas (in a Discord server of band kids, that is...) They came up with fairly convincing themes and musical selections with fleshed out stories.

The thing is, all of the show ideas had obvious stories or overdone themes... Do shows that tell overt stories ever win? I feel like most successful shows guide the audience with a theme but leave parts up to interpretation.

Another problem is that the musical selections are largely chosen by title rather than content. An AI-generated show about the seasons had four movements for each season, and the AI included that Vivaldi concerto. Not only is the piece overused, but it shows how little "thought" (AIs can't think, but I can't think of a better word) was put into the selections. Surely Vivaldi's The Four Seasons isn't the only piece that conveys the different aspects of the four seasons. Also from what I've seen, ChatGPT has a limited knowledge of how a piece actually sounds and can merely use a few adjectives to describe them. It can combine intense pieces for an intense movement without much regard for the symbolism of the source music. The above seasons show used "Summer" from The Four Seasons... in the winter movement.


AI composing music is a whole other thing. I'm not sure how AI compositions sound in 2023, but the ones I've listened to sounded like sh*t. I don't think AI can generate drill right now either. And AI-designed props and silks will open the AI art debate pandora box, so I'll stop here.

Based on what I've seen, I highly doubt that AIs can match the imagination and creativity of show designers. Would an AI use a piece about time for a show about stars? Could an AI be able to come up with complex show trilogies about life and death?

[hawknate14]

Yeah for the foreseeable future I wouldnt worry on a top level because, while I could see an AI pumping out canned shows for non or lowly competitive bands in the near future, it still lacks the creativity to really be successful on a high level of designing.

[floridaman]

Unless something groundbreaking comes out in the next year or 2, I'd give it around a decade. They're still tweaking and coming out with new versions of these engines that it will take a few years for one that can truly learn to understand art in the form of music, film, animation, etc. Like you said, it's going to take a community to really fill it with enough information first. The reason you see a lot of successful AI songs created from voices like Michael Jackson, Eminem, Drake, Ariana Grande is because there is already a wealth of source material for those engines to pull from as well as armies of fans that help mold them from continuously inputting commands and uploading that material. 

Now fast forward those 10 years, you'd definitely be able get completely AI generated Michael Gaines-esque drill by uploading sheet music. I'm sure the trouble would first pose a threat to composers and drill writers in the form of not only having their permission, but seeing other new forms of licensing within the band world arise. 

Good conversation to have to keep in the back of our minds for when the opportunity finally comes. 

[josephbandfan]

Could be a really funny thread to look back on lol

[cybrunette]

Jun 6, 2023 17:31:45 GMT -6 allthingschaotic said:

Jun 6, 2023 15:46:04 GMT -6 cybrunette said:

It's only a matter of time before someone or a group of people start feeding AI enough information to reference for these AI engines to generate shows & scores that are clean enough to blend in with even the most competitive marching bands & corps. I also wouldn't find it surprising if someone decided to start a Discord server just for this purpose. 

I've already seen some people ask ChatGPT for show ideas (in a Discord server of band kids, that is...) They came up with fairly convincing themes and musical selections with fleshed out stories.

The thing is, all of the show ideas had obvious stories or overdone themes... Do shows that tell overt stories ever win? I feel like most successful shows guide the audience with a theme but leave parts up to interpretation.

Another problem is that the musical selections are largely chosen by title rather than content. An AI-generated show about the seasons had four movements for each season, and the AI included that Vivaldi concerto. Not only is the piece overused, but it shows how little "thought" (AIs can't think, but I can't think of a better word) was put into the selections. Surely Vivaldi's The Four Seasons isn't the only piece that conveys the different aspects of the four seasons. Also from what I've seen, ChatGPT has a limited knowledge of how a piece actually sounds and can merely use a few adjectives to describe them. It can combine intense pieces for an intense movement without much regard for the symbolism of the source music. The above seasons show used "Summer" from The Four Seasons... in the winter movement.


AI composing music is a whole other thing. I'm not sure how AI compositions sound in 2023, but the ones I've listened to sounded like sh*t. I don't think AI can generate drill right now either. And AI-designed props and silks will open the AI art debate pandora box, so I'll stop here.

Based on what I've seen, I highly doubt that AIs can match the imagination and creativity of show designers. Would an AI use a piece about time for a show about stars? Could an AI be able to come up with complex show trilogies about life and death?

Leave it to this current generation of band kids to make the most out of Discord lol 

Right now in 2023 I absolutely agree with you, but I don't see why AI would have any issues with portraying time and life / death in the form of music & drill as technology advances. It would also probably have a greater understanding of those concepts than people would (I don't mean greater as in better, I mean greater as in another plane of understanding). AI bases its intelligence on human knowledge / understanding (what's given to it) so while it would maintain a somewhat human response, if I were to give an AI engine that has been developed & coded a little further a prompt to expand on specific concepts we could see something with more depth. 

----

Just wanted to throw this out there too... My thoughts in creating this thread were based on what the next shift in marching band is going to be. In the early / mid 10's when all the movement, choreography & electronics started growing into the foundation of what shows were becoming there was a clear division in how people felt about that. Sure there were those that didn't care, but back then you hated it or were open to seeing what it would evolve into. Sure you still have those that want nothing to do with the current trends & have everything more traditional, but it's finally come to a time where it has become something the large majority have taken a liking to.

While I think there will definitely be other forms of evolution before this comes into play, I have a feeling using AI will define a new generation of marching band. It'll be the next time we see a substantial divide in our thoughts & responses to change. The dynamic being one end of the spectrum "That band is using AI" or "That's not real music" to the other end of "now this level of design & composition is accessible to everyone" or "we can save a bit of money here and there".

[supersound]

I genuinely do not believe this conversation is constructive or necessary. AI will not get to a point it can churn out meaningful designs on a “new plane of understanding”

I really don’t believe that “ using AI will define a new generation of marching band. It'll be the next time we see a substantial divide in our thoughts & responses to change. ”

I think it’s overstated, it’s a trendy fad that’s certainly evolved the way we write, think and develop but ultimately when it comes specifically to the field of the marching arts, unless someone specifically spends the next five years meaningfully curating an AI to specifically complete this task (and I’m doubtful anyone will, and if someone does that they’ll have the capacity to bring it to full force).

I think this is overstated, and there are more important things and evolutions to the marching arts that we should focus and dedicate time to.

[cinnamonpromenade]

Jun 6, 2023 23:06:38 GMT -6 supersound said:

I genuinely do not believe this conversation is constructive or necessary. AI will not get to a point it can churn out meaningful designs on a “new plane of understanding”

I really don’t believe that “ using AI will define a new generation of marching band. It'll be the next time we see a substantial divide in our thoughts & responses to change. ”

I think it’s overstated, it’s a trendy fad that’s certainly evolved the way we write, think and develop but ultimately when it comes specifically to the field of the marching arts, unless someone specifically spends the next five years meaningfully curating an AI to specifically complete this task (and I’m doubtful anyone will, and if someone does that they’ll have the capacity to bring it to full force).

I think this is overstated, and there are more important things and evolutions to the marching arts that we should focus and dedicate time to.

What simultaneously worries and fascinates me about AI is how rapidly it's made the impossible possible. For example, it's automated activities that otherwise could only be performed by humans. Feed chat GPT some information and bam—here's an essay, here's a painting. In ten years from now, can someone make a living as a journalist or author? As a painted or graphic designer? (While the answer is probably "yes," it probably will complicate who can make it in those industries.)

AI right now only can work with what's already known. So the main advantage humans will still have over AI is our ability to have a unique point of view. AI can't invent new ways of thinking.

That being said, how do we know that AI won't influence the marching arts? Maybe AI won't ever write drill in Pyware... but surely it could design forms in motion that could be abstracted into Pyware. Why pay a drill writer $3-10k for their ideas when you could use AI to iterate ideas for you? Same thing with music composition. Same thing with show ideas!

Again, the main advantage designers will maintain is their unique point of view. But even then, perhaps we can replace these vast teams of hired designers with a single program coordinator able to manipulate AI well enough to design all parts of the show. Or hired designers become design consultants that provide initial ideas and some creative guidance, but otherwise do a fraction of the work they did before.

I agree with floridaman that it'll likely take a decade before we really see its influence on the marching arts, but I do think it's coming.

True, this conversation isn't more than just people sharing their conspiracy theories... but we need not be so haughty as to suggest this is negative or unnecessary. I'm sure folks said the same after seeing Crown choreograph every count of their 2015 show. Tell me, was that just a trendy fad—or did that open the floodgates for a new era of show design?

[cinnamonpromenade]

I'm most excited about AI's potential in growing smaller and/or underresourced programs. A lot of young folks have a high taste level but lack the connections to experienced designers that can provide great ideas or caution against common pitfalls.

To quote Ira Glass:

“All of us who do creative work get into it because you have good taste. There is stuff that you just love. But there is a gap, where for the first couple years you are making stuff, what you are making isn’t so good. It’s really not that great. It’s trying to be good, it has ambition to be good, but it’s not quite that good.”

“But your taste, the thing that got you into the game, your taste is still killer. And your taste is good enough that you can tell that what you’re making is kind of a disappointment to you, like you can tell it is still sort of crappy. A lot of people never get past that phase, a lot of people at that point, they quit.”

“The thing I would say to you with all my heart is that most everybody I know who does interesting creative work, they went through a phase of years where they had really good taste, but they could tell that what they were making wasn’t as good as they wanted it to be, it fell short. It didn’t have that special thing we wanted it to have. Everybody goes through that, and if you are going through it right now, you got to know that it’s totally normal, and the most important thing you can possibly do, is to do a lot of work, do a huge volume of work. Put yourself on a deadline.”

“It’s only by going through a volume of work that you are going to actually catch up and close that gap, and the work you are making will be as good as your ambitions.”

AI can accelerate that volume of work for young designers. By having this tool that can iterate an infinite number of ideas, it can accelerate a person towards ideas that match their taste level. Rather than needing to design, say, 20-30 shows before feeling confident in one's design skills, you may only need to design a couple. I think that's HUGE and I think this could feasibly start happening in the next few years.

[bandirectorman]

Jun 7, 2023 14:53:02 GMT -6 cinnamonpromenade said:

Jun 6, 2023 23:06:38 GMT -6 supersound said:

I genuinely do not believe this conversation is constructive or necessary. AI will not get to a point it can churn out meaningful designs on a “new plane of understanding”

I really don’t believe that “ using AI will define a new generation of marching band. It'll be the next time we see a substantial divide in our thoughts & responses to change. ”

I think it’s overstated, it’s a trendy fad that’s certainly evolved the way we write, think and develop but ultimately when it comes specifically to the field of the marching arts, unless someone specifically spends the next five years meaningfully curating an AI to specifically complete this task (and I’m doubtful anyone will, and if someone does that they’ll have the capacity to bring it to full force).

I think this is overstated, and there are more important things and evolutions to the marching arts that we should focus and dedicate time to.

What simultaneously worries and fascinates me about AI is how rapidly it's made the impossible possible. For example, it's automated activities that otherwise could only be performed by humans. Feed chat GPT some information and bam—here's an essay, here's a painting. In ten years from now, can someone make a living as a journalist or author? As a painted or graphic designer? (While the answer is probably "yes," it probably will complicate who can make it in those industries.)

AI right now only can work with what's already known. So the main advantage humans will still have over AI is our ability to have a unique point of view. AI can't invent new ways of thinking.

That being said, how do we know that AI won't influence the marching arts? Maybe AI won't ever write drill in Pyware... but surely it could design forms in motion that could be abstracted into Pyware. Why pay a drill writer $3-10k for their ideas when you could use AI to iterate ideas for you? Same thing with music composition. Same thing with show ideas!

Again, the main advantage designers will maintain is their unique point of view. But even then, perhaps we can replace these vast teams of hired designers with a single program coordinator able to manipulate AI well enough to design all parts of the show. Or hired designers become design consultants that provide initial ideas and some creative guidance, but otherwise do a fraction of the work they did before.

I agree with floridaman that it'll likely take a decade before we really see its influence on the marching arts, but I do think it's coming.

True, this conversation isn't more than just people sharing their conspiracy theories... but we need not be so haughty as to suggest this is negative or unnecessary. I'm sure folks said the same after seeing Crown choreograph every count of their 2015 show. Tell me, was that just a trendy fad—or did that open the floodgates for a new era of show design?

I have always thought that more internal staff should be writing and arranging their own shows, rather than purchasing Wes Cartwright design to appear credible. This is great news in a way. Maybe.

[floridaman]

Jun 6, 2023 23:06:38 GMT -6 supersound said:

I genuinely do not believe this conversation is constructive or necessary. AI will not get to a point it can churn out meaningful designs on a “new plane of understanding”

I really don’t believe that “ using AI will define a new generation of marching band. It'll be the next time we see a substantial divide in our thoughts & responses to change. ”

I think it’s overstated, it’s a trendy fad that’s certainly evolved the way we write, think and develop but ultimately when it comes specifically to the field of the marching arts, unless someone specifically spends the next five years meaningfully curating an AI to specifically complete this task (and I’m doubtful anyone will, and if someone does that they’ll have the capacity to bring it to full force).

I think this is overstated, and there are more important things and evolutions to the marching arts that we should focus and dedicate time to.

Don't really see the issue in this discussion. It's as harmless and curious as the BOA "What If" thread.

Will it work? Maybe it will, maybe not. It's always hard to discuss topics that are ahead of their time. But remember there's plenty of old tech that didn't break through 10 years ago that was considered a 'fad' yet today other companies picked those failed ideas up at the right time and now they're flourishing.

Nothing here is stopping anyone from thinking forward in other areas to progress the marching arts. This is only another mechanic that can be used as a tool in the future be it if anything even comes out of it. But putting an unnecessary wall up towards those with fresh ideas is a little much. Marching arts will always evolve and I and everyone else here has the ability to contribute in more than one area. We don't have tunnel vision.

There are millions of people around the world with some sort of musical background and millions more to come. It'd be VERY weird to think there won't be a group of them out there that will ignore this just because there are other things to work on right here, right now. Seeing what a lot of people in this very forum do with their free time by putting together collections of information to make it easier for others to access is just as viable as someone working in an AI engine on their free time.

I personally think it's great to see people creating a spark and I can only hope we see more ideas come as we move forward.

[boahistorybuff]

This is a very interesting topic. I am also somewhat eager to see if this does become a thing. 30 years ago there were many who said that in the future shows would be filled with singing, dancing, electronics and wild costuming. Many, including myself, did not think that shows would dare go too far into that stuff, but here we are. I would not at all be surprised to see AI revolutionize the look and sound of band shows. Time will only tell.

[philodemus]

For me, there are two great pleasures in watching a show: being amazed by the skill of the performers, and being amazed by the creativity and craft of the designers. Nothing makes me happier than seeing something truly original, performed at a high level, and feeing amazement at the other human mind that created this wonderful, audiovisual moment in time.

So, AI show design would rob me of approximately 50% of the joy of spectating.

I really, really hope we don’t go down this road.

But… we’re going to, aren’t we?

[ilikeguard]

I’ve gotten to a point in my life where I’m able to write professionally now. I see a lot of worrying from hobbyists that AI is going to take writing opportunities away, but this isn’t a concern you see very much from professional writers. AI can make a lot of content in a very short amount of time; this is fine for automating mindless content, but has no place in genuine intellectual or creative spheres. It’s incredibly obvious when something is written by AI. It’s formulaic and uninspired; no amount of randomization and fluff can disguise a lack of intention. I don’t fear AI coming for my craft whatsoever, and whoever feels as though they cannot compete against it really needs to find a skill that they feel more confident in.

I believe this is how the prospect of AI show design will pan out. Lots of people will make lots of shows using AI. At the end of the day, those that are purely AI-generated will be obvious (and rife with plagiarism). In my opinion, a far more likely implementation will be the use of AI as a rough draft that is then refined by a human. It would be nice to use AI to look at the best possible color combinations for visuals and to look at pieces that share similar key signatures or tempos for smooth integration in arranging – I think these are perfectly acceptable uses and will only help to make the human side of show design a bit less intensive.

Remember, the camera didn’t kill painters. It made painting more competitive, but people still paint. It just created a cheaper, faster alternative. And it’s very hard to sell a photograph as a painting, especially to an intentional eye.

[philodemus]

A random thought: is it possible drill design will be saved from AI simply by being niche and small in scale?

As I understand it [based on pop science podcasts and articles, so take that with a grain of salt] all of these 'large language model' AI basically start by analyzing as many data points as they can find. This is no trouble for something like, say, research papers, because there are, you know, a ton of research papers available online... same for news articles or photographs or portrait paintings, etc. The model says, "Well, in these other 10 million examples, the most likely word to come next is 'rutabaga' so I'll write that word here."

But with drill writing, it seems like that first step of analysis will be really hard.

First, have there even been enough drills written in the history of mankind for a 'large language model' to work?

Second, would an AI have access to those drills? It's not like there are giant, open source Pyware archives online... are there?

Now music, sure, that's already done. Graphic design for props, flags, etc.? Sure, already done. But drill, specifically... maybe safe for now?

Or maybe I just don't understand. This is often the case.

[Samuel Culper]

The AI bots are reading this thread and redoubling their efforts to improve and create shows that prove wrong the doubters.

When it happens, you responsible ones know who you are!