CULTURE CURRENT-Piotr Mirowski on the promise and perils of AI in theatre

Piotr Mirowski, a Poland-born computer scientist and actor, combines theatre improvisation with AI research, exploring the intersection of technology and human performance in his latest production, "Improbotics Presents: Artificial Reality."

Piotr Mirowski occupies a rare space where theatre improvisation overlaps with artificial intelligence research. The Poland-born computer scientist and actor has spent years experimenting with the ways that performers, audiences and language models ​can interact, producing hundreds of live shows and a body of academic research. Now a visiting researcher at Goldsmiths, University of London, Mirowski is preparing to take his latest production, “Improbotics ​Presents: Artificial Reality,” a high-tech satire in which a robot performs alongside a human cast, to this month’s Edinburgh Festival Fringe.

Speaking ‌with Reuters ​from London ahead of the show, he discusses the possibilities and risks of AI in theatre — and why technological advances may ultimately deepen our appreciation of what makes us human. This conversation has been edited for length and clarity. Were you an actor first or a scientist first?

I started doing improv comedy during my computer science studies in France. And I continued it in an amateur but serious way when I was in the United States. When I landed in the UK over 13 years ago, I enrolled part-time at the London School of Dramatic Art. How did you connect ‌acting and AI?

The intuition to me came by thinking about how hard it is for human actors to be spontaneous. Something that children do very easily, but adults have trouble doing is to literally say the first thing that comes to your mind without editing and self-censorship; thinking that's silly, that's stupid, I'm not going to say that. I was working on statistical models that were doing next word prediction — essentially what we call today language models and those models were doing it without hesitation. I noticed that the machines were able to do something that humans couldn't easily do. And so that observation pushed me to try to bring a language model on stage.

I decided to try to improvise alongside my own language model. At the time there were very few around, so I just had to design one from ‌scratch and I trained it on subtitles and movie dialogue and I brought it on the stage of a small theatre in London. How did you come to set up the theatre company Improbotics?

I learned about Kory Mathewson, who was doing a PhD in robotics at the University of Alberta in Edmonton in Canada and who had a similar idea. I wrote him ‌a very long email, and a few days later we met in London. We started working together and we created our theatre company. I have seen theatre shows also addressing AI. How does your show fit in? I had a feeling that the novelty (of AI) was wearing off and tiredness was settling in. And that's why we decided to address our concerns in this brand new show called “Improbotics Presents: Artificial Reality.” It's a parody of a reality TV show where contestants are eliminated by the audience to be replaced by an AI version of themselves or by an AI celebrity.

We blend the human creativity of improvisers and the weirdness of machine intelligence to respond to uncanny times with dark comedy. This show is effectively a satire of capitalism, as well as of our appetite for replacing ourselves whenever we feel it’s convenient.

How big is the cast? We are actively making choices in the production of a show to involve as many humans as possible. Currently (there are) 11 of us. We ⁠are also getting a ​full-time person to operate the AI. For us, adding AI means adding a human.

Is your show supportive ⁠of or against the use of AI in theatre? Our show is neither supportive nor against. There is nowadays a very binary perception of a conflict between “luddite” creative artists and “philistine” technologists. This opposition forgets the nuance that artists love tinkering with tools that advance their craft, and that technological improvement requires various forms of creativity.

In general, I believe AI should be seen as an enabler for something that the artist would not be able to do without that new tool. What do ⁠you think should always remain human?

We believe that audiences are craving genuine, human experience. The show was written by a writer's room of four professional actors, comedians and designers who felt strongly about the rise of AI in show business. We have not used AI in writing the show structure.

Is AI a threat? I am extremely concerned about the way AI is used and deployed. There is a bit of optimism in the sense that we (Improbotics) believe ​it can be harnessed.

Most of the company is made of professional actors. We all design and co-design the shows and we try to find games that are enjoyable for the audience and enjoyable for ourselves. And there is a sense of agency and ownership of the process. I think there are two things that are happening at the ⁠same time: The absolutely amazing advance of AI, (which) is the result of millennia of human work and human civilisation. But at the same time, we gradually realise that AI can produce words; but that doesn't mean it produces meaning.

We suddenly become more attuned to understanding what we actually want out of writing. And it's not just words. How about the other existential threat — to the climate?

Many of us at Improbotics have very strong concerns about the impact of AI on climate change, on human cognition, on societal ⁠radicalisation ​and on jobs in the creative industries, and wanted to use theatre and various AI tools to talk about some of these concerns. I believe that the current scale of deployment of AI across all industries can add to the existential threat of climate change because of increasing energy consumption due to the sheer scale of AI usage. This should prompt everyone to use AI intentionally and only when needed.

However, individual usage of AI at the scale of a theatre production will likely be negligible. What theatre can contribute to is to the dialogue on individual usage choices, and on structures of power and of market forces that drive the massive adoption of AI in every aspect of human activity.

How do audiences respond to your shows? When we did the Edinburgh Fringe in 2023, we ⁠ran surveys.

Our main finding was that after watching our show, audiences felt more enthusiastic about the use of AI as a tool for supporting human creativity, but less positive about AI being an autonomous storyteller by itself — a result that we found encouraging as technologists and reassuring as artists. How do you deal with the fear of AI?

We’re trying to convey the ⁠message that it's up to the humans who deploy it to make it a tool and not a replacement ⁠and that is a responsibility that is shared by all of us. We should as a society educate ourselves to value work and individuality as opposed to something that is only mass produced or synthetic and synthetically generated.

There is a realisation of the value of the shared human experience. I am positive because I know that Edinburgh Fringe has 4,000 shows and many are talking about the same things and yet the audiences still go to see these shows because they are going to see a specific performance of a specific person telling their own story. We want to ‌see a fellow human being struggle on stage and cheer us up ‌and amaze us with their own story. I hope that we're going to witness the resurgence of the author in this new age. The perspectives expressed in Culture Current are the subject’s own and ​do not necessarily reflect the views of Reuters News. (Editing by Yasmeen Serhan and Aurora Ellis)