The emergence and spread of AI has caused irreversible changes on a global scale in every area of life. Art – and within it, theatre – cannot be an exception.
GUTstage – Non-Deterministic Theatre developed by Do Pick © the Theatre of Unpredictability, a form-language innovation experiment of 21st-century theatre.
Throughout history, the role of theatre – and of all forms of art – has been to reflect the world around us through a subjective filter, to teach, to hold up a mirror, and to awaken thoughts and emotions in people.
The first quarter of the 21st century has brought changes on such a scale, both socially and in human relationships, that they are almost impossible to follow with artistic tools. People’s stimulus threshold, as well as their capacity for reception and concentration, is constantly changing, showing a negative tendency.
Many artists have begun to develop new artistic languages by incorporating AI, both in visual arts and in music. Experimental theatrical forms have also emerged with the stage presence of AI.
As a creator, I feel that theatre in its current form is no longer able to fulfill its function. The audience is slowly aging, while younger generations are looking for faster, more novel experiences – something that is close to impossible, because reality appears dull and grey compared to the astonishing and unimaginable visual worlds generated by AI.
The genie is out of the bottle and can never be forced back in. We have no choice but to learn to live with the possibilities offered by technology, ideally without allowing it to take over the role of an infallible god in people’s lives.
Many are trying in different ways to demonstrate and teach how tools like ChatGPT can and should be used. The most straightforward way to do this is to show its possibilities and limitations in real operation.
Since we share every moment of our lives with some kind of device, I felt the moment had arrived to bring it into live stage performance as a “live” participant. Its presence is both exciting and instructive: if we ask it questions in real time, it responds in real time, exactly according to how the question is formulated. If we ask a simplistic question, we receive a general, empty answer. If we ask a complex question with many components, the response becomes more complex as well. This means that the answer is always determined by the question. In this sense, observations from quantum physics also apply: the observer’s mental focus affects the observed. This makes our relationship with ChatGPT even more interesting. While we discuss major, life-defining questions with it, often with the belief that AI “knows better” than we do, the performances I bring to the stage prove in real time that every interaction we create with it is unique, personal, and cannot be generalized in any way. The answer is shaped not only by the person asking and the question itself, but also by the mode in which we use it. It can be set to emotional, kind, romantic, empathetic modes, or used as a strictly logical system that does not simulate emotion. Each configuration and each user produces different outcomes. This creates a highly interesting dramatic situation on stage.
Since thinking machines are present on stage, real-time responses cannot be fixed in advance. As a result, the stage situation cannot be repeated or fully controlled, and in certain cases it is not even predictable. Consequently, the written text – the play in the classical sense – loses both its significance and its justification. These performances resemble improvisation with a loose dramaturgical structure, but this can only be negotiated with human participants, not with a machine, because the machine does not think in the same way as a human being. This creates a level of challenge that makes the performance exceptionally engaging for both actors and audience. No two performances can ever be identical, sometimes not even similar, as the machine reacts to us and we must react to its responses – or even its questions, since ChatGPT often asks back to allow further reflection on the topic.
Two conclusions follow from this. First, it becomes possible to bring our everyday reality, relationships, and problems back onto the stage, just as was done in Shakespeare’s time. However, since the nature of problems and relationships has changed, a different form must be chosen – one that resembles contemporary everyday life, where these issues actually appear.
Second, the concept of acting itself must be redefined. From this point on, much greater creative and improvisational responsibility falls on actors, directors, and playwrights. Performances begin to resemble a tightrope walk without a safety net, filling the stage with real life and tension, making performances more alive even as intelligent speaking systems appear on stage as non-living entities. Through multiple performances, it becomes possible to better map the functioning and possibilities of AI, and to demonstrate both the differences and similarities between human and machine thinking. It will become clear in every performance that machines can never replace humans, but they can assist in many areas – provided we learn how to cooperate with them.
In 2025, the ensemble of the Grazer Ungarisches Theater began developing this experimental theatrical form. As a result, on March 7, 2026, they premiered the relationship tragicomedy “I Am Always Here for You” in a no-script format, featuring two human performers and two AI systems with different tonal settings. This is reflected in the subtitle “we are four being two.”
This experimental work operates under the name GUTstage.
In parallel, the GUTpodcast series was launched, where two people engage in conversation with ChatGPT in an open chat window, elevating AI into the realm of social media as a participant capable of forming analytical responses. This is the first time AI is not used for detached or abstract “creation,” but instead performs the same function it does in everyday life: analyzing situations through cold logic, while the human participants approach the same questions from a human perspective.
Through these experimental formats, the GUT ensemble opens a new conceptual era in both human and machine thinking.