Musical Theater in the Age of AI: Part I

Abundance and Happiness: An AI paintbrush amplifies the voice of a refugee artist

Rebecca Burnham: If you've been here awhile, you may already read about Paul Anderson when I wrote about his production philosophy that prioritizes people. Today, he returns not as a guest, but as part of the team that is building Summit Stages into an exciting business. You'll have an opportunity to meet team member Michelle Linford in the near future.

Paul Anderson: When Rebecca first invited me to write a post about “musical theater in the age of AI” (with only a few days lead time) I was initially hesitant. There is so much that could be said on the subject, and much that IS currently being said about AI generally in the world right now, and about AI and the arts generally. However, I do feel that I have a few valuable things to say on the subject, and I hope that at least some part of them will be helpful to you!

I have decided to break my thoughts into two articles on the following topics:

1. Being a Creator in the Age of AI (Agency and Creativity)

2. Performing Arts in the Age of AI (Transcendence, Community, and Corporeal Experience)

I hope that something shared in the following parts of this multi-part article will be of help to you in some way.

Being a Creator in the Age of AI (Being Agents that Act)

A Historical Event

In late Fall, 1938, a strange, meteorite-like object landed in Grovers Mill, New Jersey. Experts were immediately dispatched to examine the object, along with a news reporter who broadcast the investigation live on the air. To the public’s shock, a large, monster-like creature with several tentacles emerged from the object. Police officers, waving a flag of truce, approached the creature, but the creature fired some sort of “heat ray”, incinerating them and setting the nearby trees ablaze as the crowds scattered in chaos. The radio broadcast was cut short apparently due to difficulties with the transmission. 

As panic ensued, “emergency response bulletins [gave] way to damage and evacuation reports as thousands of refugees [clogged] the highways,” according to the Wikipedia page on the event. Phone lines and switchboards jammed with panicked callers and many others began fleeing their homes or to defend them against the Martians who were invading the earth. Nation-wide hysteria ensued. 

Orson Welles during the broadcast, and an example of the headlines that followed

Obviously, Martians did not actually invade the earth in 1938, but what you may be surprised to learn is that part of the above description actually did happen as part of the “event” – phone switchboards were indeed overwhelmed with people calling to verify whether the “news broadcast” being broadcast over the radio was real or not. In reality, it was Halloween night and Orson Welles on The Mercury Theatre on the Air CBS radio program had decided to do a radio drama adaptation of H. G. Wells' The War of the Worlds, which apparently had such high fidelity (see below) to an actual news bulletin radio broadcast, that many people were at least confused if not sent into a panic, convinced Martians were actually invading the earth. While the exact extent of the “mass hysteria” is debated, it’s clear that at least a portion of listeners misunderstood and may have even fled their homes or the state as mentioned above. (You can listen to the full broadcast yourself here, if you are interested).

Two Questions about Your Use of Technology

The point of this story is not to prove how powerful the media can be, but rather to illustrate an important principle that has applied to all new mediums and technologies since history began – from the invention of fire, to premodern weapons, to the printing press, to handguns, to engines, to radio, to television, to cable TV, to the internet. Whenever a new technology (including media technology) comes on the scene, there are two big questions we need to ask ourselves (I’d cite my source, but I’m the source and this is the first time I’ve published them, so this is the source 😊):

1. Will you use the new technology as a tool to act or will you be acted upon by it?

2. If choosing to act, will you use it to act upon others or to help empower others to act more themselves?

As humans, we all have agency – that is we are agents who can choose to “act for ourselves” or to be “acted upon” by outside forces. And despite a certain Winkie prince’s initial musical opinion otherwise, life is typically LESS painless for the brainless. As British Prime Minister Benjamin Disraeli put it in the mid-1800’s, “Man is not the creature of circumstances. Circumstances are the creatures of men. We are free agents, and man is more powerful than matter.” He continued with a comment that sounds strikingly like the world of algorithms and AI that we live in, “The invention of cunning deceits may, and in most instances does, prevent man from exercising his own powers.” But the good news I would add, is that does not have to be the case! We can choose to be agents that act (and create) rather than choosing to be acted up on by influences that would “prevent [us] from exercising [our] own powers.”¹ 

Clearly, the listeners to the 1938 radio broadcast who tuned in mid-broadcast (apparently missing the disclaimer that it was a dramatic presentation) and began fleeing their homes chose to be acted upon by the broadcast. In contrast, the listeners who heard the broadcast but called in to verify whether it was fictitious or not chose to act (which appears to have actually been most listeners, despite reports at the time). Another amazing example of using the tool of radio to act from the era is President Franklin Delano Roosevelt’s incredible use of the radio to talk directly to the American people which led to the development of parasocial relationships² with them in which many citizens felt they had a real, personal relationship of trust with the President.

Note that the two primary questions to ask yourself about new technology above don’t change over time. They apply to any new technology equally! And there have also been practical and moral dilemmas with the advent of any new technology over time, which humanity typically grapples with and then navigates. What does change over time, however, tends to be technologies’ potency, rate of change, and (in the case of media technology) fidelity. 

Potency is simply the “power” of the new technology (a rather vague concept). This often equates with its reach or effectiveness. It relates somewhat to fidelity. For example, there was clearly greater potency in the invention of the handgun as opposed to a bow and arrow. Some would argue that technological changes actually don’t always improve things, depending on one’s point of view. For example, some people still consider vinyl records and actual film to be superior mediums to digital media formats – and in sound and picture quality they often are! But for the values of convenience or efficiency, digital mediums would probably outweigh them.

Rate of change is simply the velocity at which new technologies develop (or regress). This has tended to accelerate over human history, with a huge spike in the last two centuries of industrialization (the mechanization of labor) and an even bigger spike in the last few decades in the age of computer and AI technologies (the increasing mechanization of thought).

Fidelity (in the media sense) is how accurately a copy (audio, video, adaptation) reproduces or represents an original. For example, a sound recording with high fidelity (hi-fi) closely resembles what the original source sounded like (such as the human voice). In a more general sense, you could say that dialogue in a novel has a high degree of fidelity if it sounds like how people would actually talk, or a monster represented in a movie has high fidelity simply because it looks photo realistic, even if no such monster actually exists (sort of like that Martian in the War of the Worlds broadcast…). Deep fakes in the age of AI have a high degree of “fidelity” to videos of the actual person they are meant to imitate (though unfortunately often very little fidelity to the person’s real character). 

AI and Being Agents to Act

In other words, although AI could perhaps be compared to the invention of the automatic rifle in contrast with a bow and arrow in potency, the principles remain the same… It is simply a technology that has rapidly increased in potency (which can be used for worse or better, it just depends how we use it) and that has especially increased in fidelity to human thought, emotions, and creativity. However, the two questions to ask ourselves about our use of AI both as people and as creators remain the same:

1. Will I use AI as a tool to act or will I be acted upon by it?

2. If choosing to act, will I use it to act upon others or to help empower others to act more themselves?

Of course, the other big difference is that we have now created a technology that doesn’t just have a “man behind the curtain” “pulling the strings” trying to act upon us, but one that can autonomously “ask” us itself (with a high degree of human fidelity that was not explicitly programmed into it by a human), “Can I act upon you?” in probably more enticing ways, tailored exactly to what it “learns” from our interactions with it. But it is still our choice what we answer!

What is AI?

An important step to understanding how AI interacts with musical theater is to understand what AI actually even is. Wikipedia defines AI as “the capability of computational systems to perform tasks typically associated with human intelligence, such as learning, reasoning, problem-solving, perception, and decision-making.” In other words, AI is simply and only mathematical algorithms finding and analyzing patterns in huge sets of data. 

It’s math, and 1’s and 0’s, and in the end nothing more than that –– there is no sentient being on the other end. It cannot actually create (in the deep human sense of that word) only generate from existing patterns, it can’t imagine, it can’t feel, it can’t discern, it can’t be empathetic or loving (or hating, for that matter), it can’t be creative. In some senses it actually can’t be generative, only regenerative. It’s simply a nonsentient, emotionless data processing machine, or in some cases whole data centers of machines (no matter how much it may be able to mimic human language describing the opposite). It can search and find patterns in large data sets (which is a whole nother topic beyond the scope of this article, about what data set an AI bot is being trained on and how that affects its output). One other important thing to note is that AI does what it was programmed to do – that means whoever programs it has great control over what it outputs, even if it appears to be “generative”. Since AI has no value system in and of itself, it simply adopts whatever values were programmed into it. If its purpose is to make money for its owners as efficiently as possible/create as much engagement as possible, it will inevitably drift towards treating people like objects. If it is programmed to assist people in giving them more desirable resources to then choose to act upon, it can be an amazing tool for helping them act. There are of course huge implications for the artistic and creative industries on intellectual property and how those should or should not be used to train AI models and how attribution and royalties should work in such cases. (As a content creator, that is a topic I am also very interested in – the legal system is currently figuring out exactly how that will work, and until then, things are very up in the air and in somewhat of a state of limbo.)

Some people are surprised to learn that AI has actually been around for decades. For example, if you used spellcheck in years past, that was an early form of AI! And what a helpful tool for work efficiency (or a non-helpful tool for helping people actually learn how to spell, depending on your point of view)! Voice-to-text, facial recognition, and text to speech are also all obviously forms of AI.

What most people probably think of as “AI” however is what is called “generative AI” which the majority of people first became aware of with the launch of ChatGPT in late 2022. Tools like Midjourney (for generating visual images) were launched around the same time, and now there are countless AI content generators, including SunoAI that generates music (which Rebecca uses regularly to create orchestration and demo tracks of songs she has written herself, as described in her article last summer). Although not directly musical theater, major players in the family-friendly and entertainment industries have just recently taken big steps to get into the game as it relates to AI, such as Disney’s $1 billion investment into OpenAI (ChatGPT’s parent company) and the current bidding war between Netflix and Paramount over the Warner Brothers buyout (the WB catalog includes countless beloved films, characters, and other intellectual property – including many beloved musicals – that could be yet another revenue source in the very near future it it is licensed to AI companies to train their AI “creative” bots). 

Being a Creator, not just a User

So what is to be done? 

While all of the above are obviously general principles that certainly apply to use of technologies and AI beyond just as musical theater creators, they provide very important context for understanding how we can use AI effectively as creators generally, and as creators of musical theater specifically!

The two overarching takeaways for me as an individual creator are: 

1. AI can never truly create, it can only regurgitate and reorganize what is given to it. While this can be an incredibly useful tool (including in expediting many of the parts of creating art, such as historical research or cutting costs for artistic scaffolding), even if artists are using AI as a “paintbrush” (such as this inspiring story told by the Havenlight Art Gallery owner of a young man named Manzi from an African refugee camp using Midjourney in just this way to create this beautiful piece of art), the best art that resonates the most with humans will always require the human touch of a human creator. And that’s wonderful news for us as creators! We’ll discuss transcendence more next week, but it certainly applies to the creative process as creators as well. 

2. We can be creators that choose to use all the tools at our disposal to act, including ways that scaffold and empower our creativity and learning (such as as an ultra powerful search engine for gathering sources for deeper research, or as tools to scaffold our creative process). Conversely, we should avoid using tools in ways that act as a crutch to our creativity. A good rule of thumb is that all creations should start and end with human touch (even if various tools, including AI, are used between the touches – this also bodes well for the current state of copyright, in which the AI generated portion of a work cannot be copyrighted (as of now)). Another good rule of thumb is to never use AI (or anything else) to avoid thinking or stretching for yourself. Instead, use it as a tool to be more efficient and informed in YOUR decision making. 

For example, Rebecca has used Suno AI to practice writing songs and to share dozens of songs with friends and family in a way she would not otherwise be able to do – and is blessing her life and many others in doing so! I personally dabbled slightly with SunoAI and determined that for me personally right now, using SunoAI to generate orchestrations for my songs I’m writing would be a crutch that would hinder me choosing to learn how to write orchestration myself (which I’ve always wanted to do)... It has inspired me to start buckling down and actually studying that for myself (not to mention that it would have been extremely frustrating to not be able to adjust little details easily the way I like too 😊). At the same time, I am using a musical notation software with somewhat realistic playback of instruments to help me practice orchestrating when I do not have an actual live orchestra at my disposal to play my music back to me to hear how it sounds while I’m experimenting. If I am using it as a tool to optimize my workflow to enable me to focus more of my time on what only I can do – truly creating – that can be truly helpful! If I’m using it as a crutch so I don't have to think or create for myself, I should stop doing that. Allowing space for “guided drift” and human “imperfection” are some of the most beautiful parts of human creation, which AI will never be able to truly replicate. 

Manzi - an artist in Kakuma Refugee Camp (Kenya, Africa)

As Manzi, the young man from the African refugee camp, expressed, “I realized that I could find joy even in this place, and I wanted to show that to others. And because I didn't have the skills as an artist, I let AI be my paintbrush, my voice.” Steevun Lemon, the owner of the art gallery which now sells Manzi’s painting, has explained, “We've talked a lot about the fears of AI, but there's a different side of it. AI gives a voice to others who would not otherwise have a voice, and certainly in the space of art. I hope as we move forward, just like Pandora's Box, we realize that the last thing left in the Box was hope. And my hope is that we see AI not necessarily as something to be feared, not as something to be controlled, but as a way to connect us as human beings in a way we would not otherwise have had.”

I, Paul, hope this has been a helpful exploration of what AI is and how we can use it as a tool to empower us as individual creators. I look forward to seeing what you create with AI and without it!

And I also look forward to exploring implications of how AI may affect the musical theater industry and audience experience more specifically in Part Two of this article: “Performing Arts in the Age of AI (Transcendence, Community, and Corporeal Experience).” Until then, so long, farewell, Auf Wiedersehen! –– Paul

—–––––––

A brief background on me, since I know this is the first newsletter I have written for Summit Stages:

I have been producing “shows” since about the time I learned how to walk, first in my basement as a two year old and scaling up in productions of various kinds from there 😊I have a particular love for musical storytelling, character-driven storytelling, and experiences that are as immersive as possible. I am also incredibly passionate about a people-first approach to both musical theater, business, and life in general (Rebecca interviewed me about this last year). I have also worked in the AV and media production industry for the last several years, spent over two decades doing in-depth research on the work of Fred Rogers (who was a pioneer in leveraging his artistic medium and the cutting edge technology of his day to do good in the world), and have a degree in Communications Studies, where I focused largely on studying prosocial media & technology effects and parasocial relationships.²

–––––––––

FOOTNOTES

¹ Note that while the news often trumpets the “bad mental health effects of social media” as a blanket statement, the actual research is more nuanced: “passive” behaviors (such as endlessly scrolling) tend to correlate with the poor mental health effects, while “active” behaviors (such as intentionally using the platforms to post, comment, and interact with people you know) actually is correlated with positive mental health effects. For example, see Coyne, S. M., Rogers, A. A., Zurcher, J. D., Stockdale, L., & Booth, M. (2020). “Does time spent using social media impact mental health?: An eight year longitudinal study.” Computers in Human Behavior, 104, Article 106160, https://doi.org/10.1016/j.chb.2019.10616. 

² Parasocial relationships are the relationships we have with characters or figures in the media/arts that are not actually real human-to-human social relationships (though they can have absolutely real emotional and psychological effects). Fandom is certainly an example of parasocial relationships, as are the following and trusting of online influencers, crying over a character’s death in a book, or even developing a trusted relationship with a brand generally. Like any other technological tool, the power of parasocial relationships can be leveraged in many ways, some helpful and some not. As it relates to AI, parasocial relationships are going to become very interesting as people can very easily develop such relationships with AI chatbots and algorithms – algorithms that while appearing to have high human fidelity have no actual human being on the other end (not even a writer or celebrity that consciously created the essence of what you are developing a relationship with).

---

Thanks for your support for Summit Stages! If you liked this post, please consider sending it on to a friend.

If you are not already subscribed to my weekly newsletter, I'd be delighted if you'd do so here.

And a deeply, heartfelt thanks to those who've contributed to my tip jar Your support is greatly appreciated!