We brought together African musicians and AI - here’s what happened
- Christo Doherty
When a team at Wits University and I started organising a pilot project around AI and African Music, I didn’t quite know what to expect.
Would the teams of AI engineers and musicians find common ground? Would the results sound unique, or stereotypically manufactured? And what would AI, if anything, mean for music creation on the continent in the long run?
Now, five months into the project and after hearing the first results, I can confidently say that collaboration between AI and African music holds more promise than even I expected.
The project teams spanned the continent, from Cameroon and Ethiopia to Kenya and Nigeria. One team collected sounds from remote tribes and, together with a chain of AI models, used them to generate a live musical intelligence. Another group explored vernacular languages in an integrated AI production platform, another team created digital archives with traditional Kikuyu musicians that automatically generated transcription and attribution, to mention a few.
Up until now, experimentation in this field has often relied on commercial models of AI to make quasi-African music, while those concerned about the dangers of AI demand greater regulation to protect African music from "data colonisation". Instead, we wanted to empower African musicians to work collaboratively with African AI engineers to find real solutions to the challenges and opportunities offered by AI.
What the project really did was to promote radical interdisciplinarity - bringing artistic practice and scientific engineering together in ways that may seem unintuitive but are bound to bring about exciting results, unanticipated by either party.
One aspect that slightly surprised me was that all of the projects in the pilot chose to engage with indigenous African musics and indigenous African music makers. They showed that African music is contextually relevant, not just about sounds arranged rhythmically. It is about the social context and the social meaning of musics. The projects embraced this importance.
The kind of work explored in this project is significant because it has demonstrated five very different ways of engaging with African Indigenous music and languages that are both relevant and that are mindful of the need to create revenue models that can actually generate income for Indigenous African musicians.
We stand at a moment where we know that AI is an unstoppable reality, transforming all aspects of work, regardless of whether we're nervous about it or not. We are already seeing real examples of this in the arts, including how AI slop is flooding music streaming services. But, at the same time, we have only begun to explore how this new technology can enhance human creativity. Only brave hands-on experimentation can really expand this.
I think we're just starting to glimpse the beginnings of this future. Black musicians have always used technology in ways that are innovative and often completely different to how the inventors have envisaged their tech.
Think of the way that the drum machine and sampling technology have been used in hip-hop. Or consider the way that auto-tune, developed originally to adjust the tuning of vocals in the most vanilla way possible, has become, in the hands of black musicians, a profoundly transformative and creative instrument for manipulating the human voice.
I truly believe that African musicians should be at the forefront of how the creative arts interact with AI technologies and make them authentically their own. In fact, it is only by putting experimentation and exploration at the forefront and making this possible on the continent that we can have the seat at the table we deserve.
Professor Christo Doherty is the Acting Angela and David Fine Chair in Innovation at the Wits Innovation Centre.