The history of music is a story of tinkers who found a way to turn new technology into new tools. From the invention of brass instruments to guitar amps to synthesisers, modern day music is the product of experiments at the cutting-edge. After decades of creative stagnation, artificial intelligence represents a creative fertiliser that can help revitalise a pop music industry gone stale.

The fundamentals of songwriting have remained stubbornly fixed for decades. Rather than new musical movements, music in the 21st century has seen an explosion of sub-genres, micro-genres, and retro-revivals. These shifts are best understood as minor variations on shared blueprints, rather than anything truly original.

To be clear, there are many amazing current artists producing some of the best songs ever recorded – but the forms, the structures, the genres, and the sounds, are simply not new.

This is not a failing of creativity. Rather, it speaks to the intrinsic relationship between musicians and their instruments; artistic tools leave an indelible mark on the work produced, and it’s been an awfully long time since musicians were graced with new tools.

Arguably, music’s last truly revolutionary instrument arrived back in the 1980s, when the first consumer priced digital samplers hit the market. Within a few short years this technological innovation had become the bedrock of hip hop, spurred the creation of the EDM scene, and completely reshaped the sound of pop.

Forty years on and music is in dire need of new instruments. There is good reason to believe that artificial intelligence will provide them.

Consider the recent explosion of voice models. While the dubious practice of modelling an artist’s voice without their knowledge or permission has rightly generated controversy, the underlying technology is nothing short of incredible. Artists like Holly Herndon and Grimes have led the charge; transforming their voices into a new kind of vocal instrument and then gifting it to creative communities.

The ability to sing with someone else's voice opens up a fascinating new chapter in the history of musical borrowing. In truth, that’s just the start. The ‘timbre transfer’ techniques which underlie these voice models can be used on any sound and will eventually power instruments unlike any instrument we’ve heard before. A musician might sing through a modelled violin, or strum their guitar through a modelled choir, producing sonic hybrids at once uncannily familiar yet utterly distinct.

Then, of course, there is the generation of music via natural language text prompts. This is perhaps the area that elicits most angst amongst creatives, yet it also might be the one that spurs the most profound new creative pathways.

With generative AI as a co-pilot, musicians can venture into uncharted territory, guided by AI suggestions and influenced by patterns it uncovers from the vast corpus of musical history. In this paradigm, the musician provides feedback to refine the model’s output, iterates upon and implements the best ideas, and then assembles these sounds into a musical collage not dissimilar to the early tape-music assemblages of The Beatles — or the next big hit.

Music is in dire need of new instruments.

That said, prompting AI models via text input is likely to be a mere stepping stone towards more musically intuitive methods. We are already seeing this take shape with Meta’s MusicGen, which can be prompted by other melodies as well as the written word.

Artificial intelligence has potential even for those who see no problems with modern pop music: by making more of it. Just as digital technology previously expanded access to music production and distribution – allowing independent musicians to create albums that would previously have required a six-figure investment – generative AI has significant potential to expand access to songmaking, making the industry accessible to a far broader range of artists.

Not everyone has access to a musical instrument in their home, and many struggle to acquire the specialised knowledge of music theory and production. No longer does that rule out a career in the industry. For these large groups of people, abstracting away that technical barrier is key to unlocking the pure joys of music.

It’s easy to imagine tools that help would-be-artists generate ideas by providing a starting melody or rhythm, or tools that help flesh out and develop a simple musical idea. These are foundational pathways that lead people into the world of music and make them feel empowered to try new things. More musicians results in more music, more amazing music, and more diversity — all of which is only a good thing.

But artificial intelligence represents an even more fundamental shift in the music industry. How much is too much? Song creation could become a cheap and automateable process, with any musical style easily replicated by a few keywords and clicks in a matter of seconds. What if industry giants follow in the footsteps of Tencent Music Entertainment and pump out thousands of AI-generated tracks, some of which have garnered over 100 million streams?

In such a future, songwriters hold a set of advantages that mass automation cannot match: music gets its enduring power from expression, narrative, emotional resonance, personal perspectives, and cultural context. All of these elements are well represented in current day songwriting, but in a world of anonymously generated music on an industrial scale, they take on a renewed, indeed an essential, importance.

The culture of music has no conclusion.

Algorithms will replicate or pretend to replicate all of those elements, but music fans fall in love not only with the song but with the songwriter. If a song is made purely through prompts, listeners will want to know who did the prompting, why they chose those inputs, and how the output ultimately relates back to their lived experience. The music production process has been accelerated by machines for decades, but listeners have always cared more about the people operating them than the machines themselves.

Music is a gestalt artform, one that transcends its technical construction to become an open-ended conversation between creator and audience. As a result, algorithms plus creators will beat algorithms at scale; the tools may be new, but the desire for connection remains the same.

The culture of music has no conclusion, no final chapter, and certainly no epitaph. It is a constant evolution shaped by artistic, societal, and technological forces, generative AI merely being the latest catalyst.

Music makers of all stripes should take heart from this simple historical truism: the human capacity for creativity is boundless. Give a person stones and a stick and they’ll tap out a rhythm. Give them the latest and greatest AI tools and they’ll create something nobody’s ever heard before — and you can be sure it’ll be worth listening to.

The history of music is a story of tinkers who found a way to turn new technology into new tools.