A crayon drawing on a folded piece of paper looks like something a kid hands you at a kitchen table. At a Los Angeles theatre in 2025, audience members made drawings just like that on their way into the room. Minutes later, those same sketches were moving on stage as animated figures, part of the world the performers were acting inside [AWS].
That is roughly where live theatre and generative AI, software trained to produce new text, sound, or images from patterns it has learned, have met so far. This isn’t robots taking bows. It’s a new presence in the room that reacts, answers back, and occasionally surprises the director. The interesting part is what it does to the humans standing next to it.
A Rehearsal Room Surprise
The oldest skill in acting is listening.
You wait for the other person’s line, and your face does something honest before your brain catches up.
An AI scene partner puts that skill under fresh pressure, because nobody in the room knows what it will say. At Boston University in 2025, a 90 minute staging of Shakespeare called 『Dream』 used a tool named Random Actor, which combines motion capture, sensors that track a performer’s movement, with generated imagery [Boston]. The performer moves, the machine answers, and the answer is never quite the same twice.
For a general reader, this means the actor cannot coast on rehearsed rhythm. They have to react for real, in front of you.
How the Machine Listens
Strip away the mystique and the setup is fairly plain.
Sensors or microphones capture what a human is doing. A generative model responds within seconds, through lights, speakers, or a screen.
In Porto, EU funded researchers staged 『Re-embodied Machine』 in September 2025. A dancer wearing motion sensors moved against AI driven lighting and sound that answered her body in the moment [European].
Where the Stage Goes Next
Scripted theatre got there earlier than most people realize.
The Young Vic’s 2021 show 『AI』 used GPT-3, an earlier AI text generator, to write its script live across three performances. Prague’s Švanda Theatre presented 『AI: When a Robot Writes a Play』, a one act drama built from machine written text [Guardian].
A pattern is emerging from these experiments:
-
AI as an unpredictable partner, not a replacement performer
-
Human directors shaping structure while the machine fills in the lines
-
Audiences invited to co-create, as in the Los Angeles drawings
“The most successful theatre companies view AI as a collaborative partner.” (Stage and Cinema) [Stage and]
A Howard University review of projects including Wayne McGregor’s 『Living Archive』 and the Young Vic show reached a similar conclusion: AI tends to function alongside artists rather than in place of them [Howard]. Meanwhile, a survey led by artist Annie Dorsen with SMU DataArts is gathering data on how these tools affect working performers’ pay and prospects [Hawaii SFCA].
It was worth almost nothing until a room full of strangers watched it breathe, and then it belonged to everybody there. That’s the useful way to think about AI on stage: the machine supplies unexpected material, and the live human reaction turns it into theatre. Next time you hear about a show using AI, watch the actor’s face in the pause before the answer arrives. That breath is the part no software wrote.
🔖
- AWS on UCLA immersive theatre with generative AI, 2025
- Boston University College of Fine Arts on Dream and Random Actor
- European Commission Horizon Magazine on AI and live performance
- The Guardian guide to UK theatre and AI, September 2025
- Hawaii SFCA on Annie Dorsen and SMU DataArts performer survey
- Stage and Cinema on AI tools supporting modern theatre
- Howard University Magazine on AI in the fine and performing arts
Photo by
Photo by