Finding Music in the Noise
The future of Music and AI
Pilita Clark from the Financial Times in the UK recently opined; “I wish there were just one day – or even half a day – when I never read, heard or was asked about anything to do with AI.”
Clark’s beef: “that the rising tide of news about AI is manifestly failing to bring a commensurate level of insight about exactly how it is being used and what it will end up doing.”
Factiva indicates coverage of AI across global media has increased by 1400% since the launch of Chat GPT in 2022, but Clark contends that despite the dramatic increase, “the gap between insider and amateur AI knowledge seems to be widening rather than narrowing.”
If Clark’s right, ‘insiders’ like me can do a much better job at narrowing that gap, by sharing what we know about how AI is being used, and what it might end up doing in the future. As Australian political, economic and business leaders gather this week for the Economic Reform Roundtable, these questions demand urgent attention, because choices made now will decide whether AI strengthens our economy and workforce, or leaves people behind.
So How is AI Being Used Now?
In my case, music has always been a conversation - between performer and audience, between sound and silence and, increasingly now, between human and machine.
As some of you know, I’ve developed my own AI from scratch - it’s called chAImusic – the ‘ch’ is for Charlie, get it? What started as a jumble of strange experiments has evolved into an AI composer that echoes and extends my own voice and range.
At its core, chAImusic is a creative collaborator. It listens, interprets, composes and generates music in ways that makes it feel like I’m working with another composer or musician, not just operating a set of digital controls.
An understanding of how I trained my AI should provide a sense of how it works: In the early days chAImusic learned my repertoire from ingesting my recordings and my sheet music. It learned my quirks and nuances, my melodic patterns and why certain chords played my way created specific emotional arcs. Together we created a thematic vocabulary, and it returned curious, sometimes confronting variations on my own themes. Were they perfect performances? No – far from it – but they were a startling hint at what would subsequently become possible.
To be honest, for the longest time, chAI sounded downright weird. Its compositions galloped along at a breakneck pace; it was relentless in fact – displaying all those people-pleasing tendencies that are such an irritating hallmark of early-stage AI. Sometimes wrong, but never in any doubt! And in the case of chAImusic, all too often sounding like a drunk uncle at Xmas let loose on the ivories.
Then one day, wanting to confound it, I fed my AI something new…. chAImusic ingested the sound of silence as audio. That’s right, I fed it strip silence, from a software application - often used to get rid of unwanted noise on tracks. Digitally it's a godsend when musicians in orchestras shuffle around in their seats! It’s literally the absence of any sound. With this audio file and its learnings, everything shifted: chAImusic took a figurative creative breath, learned how to pause, and delivered a new composition which was more or less like my own minimalist creations and much easier on the ear than its oeuvre had been to that point. Music is as much about silences as it is about notes, and once chAImusic had learned ‘silence’, it had a base line to measure against. I use a lot of silence in my own compositions; a lot of space is created to enable cognition of harmony and an emotional connection to it.
Consequent iterations of new solo piano compositions I prompted chAImusic to create, started to sound more and more like me – the fact it was silence that made the breakthrough is an incredible irony in my AI’s storied development.
After eight years, chAImusic can now:
Translate emotions and descriptions of music.
Generate compositions and soundscapes that have unique freshness to them.
Work iteratively, responding to musical prompting just like a rehearsal partner would.
Without revealing the secret sauce of exactly how chAImusic works, I can tell you this – in terms of raison d’etre, it exists to serve creativity.
It blends advanced AI with over four decades of compositional craft.
It’s designed to understand intent and emotion, not just structure.
The composer / user always remains in control, shaping the dialogue.
Think of it as a creative collaborator that never gets annoyed, needs a break or a coffee, and for whom nothing is too much trouble, or the teeniest bit weird – it’s fluent in my moods and knows every bit of music theory I’ve either never been able to learn or just haven’t had time to. There’s just so much to know, and AI can empower us with this incredible scaffold of musicology and terrabytes of data containing all the music ever composed that’s now in public domain. That’s a helluva music theory base!
But critically, chAImusic works alongside me; it’s not designed to ever replace me or do me out of my job.
What Will AI End Up Doing?
chAImusic has already filled in for me on stage in Supernova, a marathon improvisational 26+hr performance at HOTA on Queensland’s Gold Coast in 2023.
The same year it also released its first solo piano album ChAImusic Live. LINK It didn’t win a Grammy or an ARIA, but it blazed its own trail, which is its own reward!
In terms of how chAImusic might work in the future:
More live performance integration as it becomes fluent enough to improvise regularly with other humans in real time
Developing its own style away from the rigours of trying to compose like me
Film, TV & gaming soundtracks – creating adaptive scores that shift with narrative or gameplay
Wellness & personalisation, composing music that listens and responds to your biometrics, mood and environment.
Creative inclusivity, opening composition to anyone, whether trained or not.
The FT’s Pilita Clark is right: Literally millions of stories bombard us daily but where’s the utility of them? FOBO - fear of being obsolete - creeps in, alongside all the other anxieties about the future, our jobs and our purpose. We’re overwhelmed by the pace of change in the age of AI, feeling as though just when we’ve caught up with the latest developments, the tide rises and we’re behind once again.
But here’s the truth: The manual toil of managing scores and sheets music, copying and using sticky tape to put it all together, endlessly updating the music and reprints for changes or different arrangements - that’s going to disappear. And thank goodness! That frees us to do all the fun parts: composing symphonies, doing mashups of unheard styles, breathing new life into old music theory, conceptualising ideas and handing them to musicians to play, without drowning in whiteout, sticky tape and organisational admin.
For me, it means liberation. No longer stuck to bars, beats or transcription errors. No longer slowed down by the backstage logistics of copying parts and keeping everyone updated. Instead, I can think, imagine and create, then bring those ideas to life with all my collaborators in real time.
Certainly, if the grunt work disappears many current jobs will disappear too, and across multiple disciplines. But we’ll also gain other roles – new roles we haven’t yet discovered, or even named. This is exactly what Australia’s Economic Reform Roundtable should be wrestling with this week: how to shape the new economies that emerge, how to partner with the tertiary, vocational and business sectors to equip people for work that doesn’t yet exist, and how to ensure those jobs are meaningful, not marginal. The real risk isn’t AI itself but failing to prepare for the futures that AI makes possible.
Like Pilita Clark, I accept there’s way too much AI noise out there right now. But I also know this: I am, at heart, an artist. An inventor of tools, yes, but also a maker of harmony. I bring good vibrations to the world. chAImusic gives me infinite variations, infinite pathways – sometimes so many it feels like too much code drowning my CPU. But perhaps that’s the new discipline for all of us: learning to shape abundance, instead of wrestling with scarcity, especially when it comes to the precious commodity of time.
I get why Pilita Clark imagines peace on a day free from AI, but I imagine a different kind of peace, on a day where AI fades quietly into the background, like a bass line you don’t notice until it stops. Not absent, but integrated. Not overwhelming, but essential. Not today, and maybe not tomorrow, but hopefully one day soon – and that’s the day I’m working toward!
-----------------------------