Plagiary
ARTIFICIAL INTELLIGENCE,REAL DANCE.
by Alisdair Macindoe
Step into Plagiary, where the Dancenorth Ensemble becomes the playthings of an AI-speaking choreographer. A computer voice issues instructions - sometimes surreal, sometimes poetic, sometimes absurd - while audiences can follow the prompts as projected text.
Created by multi-award-winning dance technologist and choreographer Alisdair Macindoe with acclaimed media artist Sam Mcgilp, Plagiary utilises a custom-designed context-free grammar (CFG) score system and contemporary Large Language Model (LLM) text generation.
Fast, thought-provoking, and unpredictable - made new each night through live interpretation.
Upcoming Performances
Gurambilbarra (Townsville City)
16 Jul 26
7:00PM
Dancenorth Theatre
17 Jul 26
1:00PM
Dancenorth Theatre
17 Jul 26
7:00PM
Dancenorth Theatre
18 Jul 26
7:00PM
Dancenorth Theatre
Credits
CREATIVE CREDITS
Concept, Direction, Choreography, Text, Coding, Sound Design, Set Design Alisdair Macindoe
Video Design, Image Design, Coding Sam Mcgilp
Software Development, Head Coder Chris Chua
Costume Design, Prop Design Andrew Treloar
Lighting Design Amelia Lever-Davidson
Performers Sabine Crompton-Ward, Tiana Lung, Aleeya McFadyen-Rew, Jag Popham, Felix Sampson, Michael Smith and Bella Hood
Acknowledgements
Plagiary is produced by Insite Arts.
This project has been assisted by the Australian Government through Creative Australia’s Playing Australia Fund.
Dancenorth Australia is assisted by the Australian Government through Creative Australia, its principal arts investment and advisory body, and the Queensland Government through Arts Queensland, and by Townsville City Council. Dancenorth is supported by the Tim Fairfax Family Foundation and receives funding from Creative Australia through the Australian Cultural Fund.
Warnings
May contain adult language, concepts, and content.
This show is appropriate for people of all ages, however, recommended for ages 15+.
Accessibility
Plagiary is a multi-sensory performance that combines written and spoken text.
Audience members are invited to wear custom glasses as an optional creative and accessibility device. These glasses use polarising lenses, which cause some visual elements to disappear or become less visible. If you already wear polarised glasses, you may notice similar effects.
For Deaf and hard of hearing access audiences, most spoken text is also presented in a readable format. In some moments, text may be intentionally unavailable when the corresponding audio is also minimal, as part of the artistic design.
For Blind and low vision access audiences, Plagiary is compatible with an audio description system. This provides spoken descriptions of AI-generated prompts. An audio-described introduction and a tactile tour are also available to support your experience - please contact us to organise either of these options.
Additional Information
For more information or to ask any questions about the show, contact us:
Phone: (07) 4772 2549
Email: admin@dancenorth.com.au
“A thrilling and often funny testament to the resilience of thehuman body in an era increasingly shaped by AI.”Elise Peyronnet, Artistic Director, Now or Never Festival
“A clever and thought-provoking exploration of how AI can expand the boundaries of creative expression while also highlighting the irreplaceable nature of human artistry.”Brendan Daynes, Dance Life Australia
"Plagiary showed that humans can respond creatively to AI, but that AI can’t yet replace human creativity.”Savannah Indigo, ARTSHUB
Q&A with the maker
Where did your interest in AI-generated dance come from?
I have been building systems that use code to generate sound and musical composition for over a decade. This interest naturally expanded into generating novel dance instructions.
I’m also fascinated by the relationship between written, spoken and physical language, and find the translation really interesting.
What were your initial enquiries and how have they changed over time?
Plagiary, and the systems that underpin it, were built over many periods of research, coding, experimentation, and play. I started in the 2010s by building Context Free Grammar (CFG) sentence generators from dance scores and then moved on to create paragraph, 'section' and 'form' generators to house those sentences. As Large Language Models (LLMs) developed in the 2020s, I began meshing the two systems together.
How do AI tools change the creative process?
AI tools don’t fundamentally change the artistic process, but they can speed up many long-established methods of art-making such as appropriation, constraint-based composition, chance operations, procedural authorship, collage, and re-framing existing material.
The “output” may arrive faster and in greater quantity, but that doesn’t automatically increase quality. It’s our role as artists to curate, refine and present material in the way that best supports our vision. Crucially, all the same critical, ethical, and moral responsibilities apply: transparency, consent, fair labour, attribution, and accountability for meaning and impact. The artist remains responsible for the work’s intentions, methods, and consequences.
What training data is used for Plagiary’s AI and where was/is it collected from?
Plagiary uses an off-the-shelf LLM component, heavily customised with my own rule sets and curated language developed from years of my own choreographic notes, prompts, rehearsal observations; and from studying the language used to write dance scores and instructional texts.
The initial LLM was pre-trained by its developers using a mixture of publicly available web text, licensed datasets and other large text corpora. But its rule sets, curated language and show structure have come from my own hand – about 10,000 lines of code and text, written personally.
Who is the creator of the work if choreography or scripts are AI-generated?
Authorship doesn’t disappear with AI; it is, in fact, exactly the same.
In most artforms the “creator” is the person or group who sets the intention, chooses the materials and methods, makes the key decisions about selection and arrangement, and takes responsibility for what the work does in the world. AI can generate content, but it doesn’t supply purpose, context, or accountability – those remain human. So the question of who made the work is answered the same way it always has been in practices that involve tools, delegation, chance, citation, or collaboration: authorship sits in the framing, the decisions, and the ownership of consequences, not simply in who (or what) produced the raw material.
How do you feel about concerns over AI ‘replacing’ traditional art forms and techniques?
I try to evaluate art less on how new or novel the art or tool being used is, and more on how it communicates something to its audience.
Humans have been using the most up-to-date tools to express ourselves creatively for ever, and I am proud to continue that tradition. I am making art about the world I am in, using the tools and objects of that world. I can not control when I was born, and feel that expressing myself in the context I exist is the most artist-led way I can work.
How does the environmental sustainability of Plagiary compare to more traditional models of dance making and touring?
I can’t quantify an exact trade-off, but my strong sense is that the touring model is where the biggest environmental difference sits. Touring a smaller core team, and working with local dancers in each city, can significantly reduce flights, accommodation, and freight – usually the dominant costs in touring emissions. In comparison, the computer used to generate and manage text for a performance is likely relatively small, especially as we’re not training models and we’re not running continuous heavy processing.
Should AI be credited as a collaborator? What about the artists it references?
Whether AI “should” be credited as a collaborator depends on how you define collaboration. Citations and references are an important topic in art. I don’t feel a need to police either for other artists. In Plagiary, I make it clear that AI is part of the work.