Do you play AI? What constitutes good practice, and can digital tools such as artificial intelligence be compared to playing a musical instrument? Is the pianist, in fact, a cyborg – or is the AI prompter a musician?
The comparison is not metaphorical flourish; it is analytically useful. Playing AI is closer to musicianship than to the use of tools.
His fingers are almost fused with the piano
In a research interview, Artur Tuznik, a pianist–composer, explained to me how the construction of the instrument, the acoustics of the room and his own curiosity about the piano contribute to the artistic process. Gradually these elements combine to create notes that become a new, unique piece of music. Music emerges from the interaction between instrument, environment and cultivated skill.
As a researcher, I do not often find myself in musical settings, but this fusion is something I can relate to. Like most people, I do not normally celebrate the mastering of digital tools, even though that too requires intuition, practice and creativity.
I experience it myself when I program and do data analysis. A small harmony emerges, along with an adrenaline rush, when I finally get a stubborn piece of code to behave. This, too, is a form of musicianship.
I also have a colleague who plays entirely by ear in certain theoretical research contexts. She can discuss and search for knowledge on the computer, while simultaneously weaving ideas and empirical material into theory, so that her brain almost intuitively and uncompromisingly improvises an elegant piece of text.
When she and other colleagues (who, to be fair, are also skilled theoretical composers) jam at the whiteboard or perform a piece of theory, it is worth listening, and it is even more rewarding to sit in on the session with your own complementary methodological experience or theoretical instrument.
This way of composing articles is rarely linear. In my experience, creativity and improvisation are more central to knowledge production than is often acknowledged. We tend to understate how much scholarship depends on cultivated improvisation.
Playing AI as a creative practice
I have another colleague who ‘plays AI’. He, too, plays fluently. When something deeper needs to be developed, ‘thinking’ is used; the prompt may be developed in Claude, then passed on to ChatGPT, which helps develop code that, through API calls, can function far more autonomously.
He talks about it with such enthusiasm that it is almost hard to understand why it sometimes ‘sounds false’ when you try to play AI yourself. Which is perhaps a bit harsh, considering that I am an AI researcher who has worked with artificial intelligence on and off since 2007. Sometimes I feel like an old music theorist trying to contribute to an ensemble with a young, vibrant musical talent who masters the tools almost intuitively. I have to practice to play (or to be?) in tune.
Together with a third ‘AI player’, we have formed an AI band, which we started when we wrote the book AI at Work, published in Danish in 2025. Sometimes we jam, and other times we refine text and ideas together. Sometimes it turns into a bit of a mess, and other times it becomes a great piece.
So, are we AI cyborgs? Is the composer a musical god? And when the AI can co-create, what does the AI-instrument – and the human-instrument – mean to the harmony?
Cyborgs, hybrids and human-AI creativity
In the early 1980s, Donna Haraway famously said, “I would rather be a cyborg than a goddess” in her much-cited Cyborg Manifesto. She argued for merging with the digital – at the time still in its early stages, long before Facebook and the like – which she saw as something that could liberate her as a human being and as a woman.
As a cyborg, one can become something else, for instance freed from the constraints of dyslexia or from limited access to production tools, and suddenly granted an immense creative freedom. Yet one might also question how much agency a cyborg truly has when caught in the grip of American tech giants and the capitalist logic embedded in their platforms. The cyborg is not free. But the alternative is not innocent either.
Hybridity does not eliminate power structures; it makes them visible. If one is locked into an idea of a particular (in Haraway’s example, a female, goddess-like) nature, it may nevertheless be preferable to use technology to free oneself from this bond and become a hybrid without a fixed essence and external norms.
The musician describes a long process of working with the piano, where one is first bound by the instrument’s . And perhaps later still, one becomes bound by it again through expectations attached to mastery. AI practice seems to follows a similar trajectory.
Perhaps the pianist is a hybrid with the piano, one colleague a hybrid with AI, and another with her theories. Each has an instrument. The musician’s hybrid form is well known and celebrated. I am not the first to be captivated when I encounter musical performance – that is, after all, one of the reasons people attend live concerts.
The researcher’s hybrid form with her theoretical instruments is also known and celebrated. But hybridity with generative AI is treated as suspect rather than skilled.
My colleague’s hybrid form with generative AI is interpreted as far more divisive than divine. It is a new instrument, and we must learn how to play it well.
Like learning a musical instrument, learning to play AI takes time. If you have heard someone play the violin for the first time, it rarely sounds quite right. Perhaps the same is true of AI, which is still a new instrument for many.
And just as Artur Tuznik is shaped by the piano, we are shaped by our instruments. As AI is becoming an instrument for the masses, used widely in both labour, education and public policy, we need to reflect on whether we are learning to play it in harmony rather than cacophony.
As we learn to work with generative AI, maybe we can be inspired by how musicians practise. And perhaps we should become more aware of how to respond to this instrument – and how it, in turn, shapes us, even as we attempt to shape it?
Sine Zambach is an author and Assistant Professor at Copenhagen Business School.
AI in Higher Education by Sine Zambach is available for £19.99 on the Bristol University Press website here.
Bristol University Press/Policy Press newsletter subscribers receive a 25% discount – sign up here.
Follow Transforming Society so we can let you know when new articles publish.
The views and opinions expressed on this blog site are solely those of the original blog post authors and other contributors. These views and opinions do not necessarily represent those of the Policy Press and/or any/all contributors to this site.
Image credit: Possessed Photography via Unsplash