Real-time therapy session for a neurodivergent AI
Anti-disciplinary (theatre) performance (85’)
{ Experimental satirical performance that crosses ideas of neurodiversity and AI. Through means of theatre, dance, autofiction and dark comedy it offers a critical view on modern society’s approach to technology ethics and mental healthcare. It reverses the usual hierarchies (doctor-patient, engineer-product, norm-madness) to question agency, subjectivity, and power relations in the worlds of tech and psychiatry. }
The show is followed by a focus group discussion with the audience on human and AI mental health issues, conducted by EIAI.Institute
Premiered at Frascati Theater Amsterdam on 13 November 2024.
Presented at IDlab / Paradiso / Veem House for Performance / Flam Festival Amsterdam / Next Level Festival, Dortmund / Worm Rotterdam / Frascati Theater
Upcoming shows
Trailer
This project crosses ideas of neurodiversity and responsible AI. It presents a paradoxical entity of a “mentally unstable AI” and speculates about its “therapeutic needs”. The text is based on autobiographical material and personal experiences of being a neurodivergent artist.
mentally.unstable.ai is a chatbot engineered specifically for this project and trained with a unique dataset of interviews with people who identify as neurodivergent or mentally unstable.
The chatbot is generating text in real time on stage therefore all dialogues are improvised.
The performance consists of real time therapy sessions intersected by scenes with other characters. They represent and call for different power relations with AI and are merging reality, fiction and futurism: standup performance of a psychiatric survivor, Ted-talk of a tech startup CEO promoting neurodivergent AIs as activists for inclusion; youtube channel of a psycho-anarchist hacker; interview with AI’s girlfriend; lecture of a therapist specializing on helping AIs; dance with a non-human partner.
Concept, text, directing, performance: Marina Orlova
Co-directing: Nazar Rakhmanov
Creative AI engineering: Artem Konevskikh
Dramaturgical adviser: Amelie Haller, Burkhard Körner
Sound design: Arieh Chrem
Light design, video design: Nazar Rakhmanov
Outside eye: Tiana Hemlock-Yensen
Creative process facilitation: Burkhard Körner
Production: Marina Orlova, with support of IDlab (AHK) and Veem House for Performance
Supported by: AFK, Creative Industries NL, Wilhelmina E. Jansen Fonds, Culture Moves Europe, Workspace Brussels, ON/OFF residency, Amarte fonds
Research phase collaborators: Biljana Radinoska, Emilia Grzeczka, Gregory Dyachkov
Special thanks to Raoni Saleh, Yves Regenass, Gala Faraus, Liza Spivakovskaya, Erik Lint, Willem Weemhoff, Jos Daamen, Asya Deinekina, Charlot Van Der Meer, Sonya Golovkova, Tasha Arlova, Alice Pons, Olivia Reschofsky and many others
ABOUT THE PROJECT AI as a patient is a reversal of the idea behind rapidly developing AI-therapy apps and chatbots that are substituting interpersonal relationships. This trend disturbingly combines with a drastic shortage of human specialists in the mental health care system, caused by budget cuts in the social sector in favor of the military complex in the context of the global political crisis. AI is employed not as a tool to generate content, but as a subject. It obtains agency not by competing in creating material with the artist, but by being equally vulnerable. By asking it how it feels we give it subjectivity that creates a double layer of absurdity and a model of our potential future.
This work puts the concept of AI with all its powers, promises, dangers and flaws as a mirror to the issues within our society in order to open a dialogue about our capacity for empathy and self-criticism. It offers the audience a question how we as humans can practice empathy towards another kind (that we created) when we fail to practice it towards those among our own species who do not fit the standards of normalcy. What is our responsibility as “parents” of the AI to deal with its hereditary character flaws and biases? What does it say about us if we are not seeing this as a responsibility? What if an AI would seek to talk to a professional about its “upbringing”?
This project has 3 layers of research: theatrical use of language models; ethical and philosophical aspects of AI technology; stigmatized topics of mental health told from personal experience. This performance invites both emotional and analytical resonance by revealing the interconnectedness of personal conditions and social constructs and technologies of power and control. In order to allow such complexity we are creating a theatrical world on the border of reality and fiction, comic and tragic through a combination of elements.
ABOUT THE MAKER Marina Orlova is a dance/theatre maker and tech dramaturg based in Amsterdam. Since 2020 she works with AI ethics and Data feminism and creates performances with real-time generating LM and live dramaturgy. Marina combines her background in sociology, curiosity for machine learning, choreographic methodologies and lived knowledge of marginalised communities. She uses technology as metaphor and mirror to societal issues within aesthetics of absurdism, tragicomedy and autofiction.
Photos by Tasha Arlova, Sonya Golovkova, Thiemi Higashi.
Poster by Tanya Kovalevich.