African musicians and technologists are racing against time to secure their place in the artificial intelligence revolution, with a groundbreaking Wits University project demonstrating both the peril and promise of generative AI for the continent’s creative industries. The gap between Africa’s rich musical heritage and its representation in AI training data has become a critical concern, as Charles Goldstuck, founder of GoldState Music and a prominent voice in South Africa’s music technology sector, warns that without swift action, the continent risks becoming a passive consumer rather than an active shaper of AI music tools.
Speaking at a showcase hosted jointly by the Wits Innovation Centre and the Wits Mind Institute last week, Goldstuck painted a stark picture of how African creators are being left behind in the global AI music arms race. The event marked the conclusion of a significant six-month pilot initiative—the AI and African Music Project—which paired musicians across the continent with AI engineers to develop technology specifically designed for Africa’s creative sectors. It’s an effort that underscores just how urgent the need has become to build African-led solutions before international players cement their dominance.
The challenge facing our continent is multifaceted and accelerating. Globally, the commercial development of AI music platforms has vastly outpaced regulatory frameworks, meaning that corporate agreements and licensing deals are now shaping the industry’s future far more than legislation. Suno, valued at US$2.45 billion following a massive $250-million funding round in November 2025, is currently locked in copyright litigation with major rights holders, relying on a “fair use” defence. Meanwhile, Udio has taken a different route entirely, securing licensing agreements with Universal Music Group, Warner Music Group, Merlin, and Kobalt—essentially building what Goldstuck describes as a “walled garden” approach.
What makes this landscape particularly troubling for African artists is the sheer scale of AI-generated content flooding streaming platforms. Spotify removed more than 75 million spammy tracks from its platform over the past year, yet the problem continues to grow exponentially elsewhere. Deezer has disclosed that it receives over 60 000 fully AI-generated tracks daily, with an estimated 85% of streams on AI music being fraudulent. These platforms are now forced to demonetise and exclude AI-generated music from royalty pools, a move that protects human artists but also reveals the chaos of the current system.
One of the most pressing technical challenges remains largely unresolved: attribution. Even as detection tools for AI-generated music become more sophisticated, tracing an AI composition back to the specific copyrighted works used to train it remains in its infancy. Goldstuck highlighted that current compensation models often rely on general market share rather than precise usage data, which means artists have little visibility into how their work is being used to train these systems. Whether attribution technology can be effectively developed at scale remains an open question.
Why Africa needs AI music tools built for African creators
The real concern, however, extends beyond copyright disputes and fraudulent streams. Africa’s cultural contribution to global music is fundamentally misrepresented—or entirely absent—from the datasets being used to train these systems. Most AI music platforms have focused their development efforts on Anglo-American and Chinese repertoires, meaning African genres, languages, and musical traditions have been largely overlooked during the initial wave of AI training. While this has spared the continent from some of the worst impacts so far, Goldstuck warned that Africa currently lacks the technical infrastructure, funding, and decision-making power to influence how AI music technology develops globally.
This is why the Wits project matters so much. Rather than waiting for international companies to build tools that might—or might not—suit African creators, the initiative brought together technologists and musicians from across the continent to develop solutions rooted in local contexts and needs. The showcase featured five compelling prototypes developed by teams spanning seven African countries, each addressing different aspects of how AI can serve African music communities.
Zazi, created by South African artist Umlilo and Ghanaian AI engineer Gideon Gyimah, demonstrates a “musical digital twin” that enables real-time interaction with voice, rhythm, and storytelling. The Bɛ̀bɛ̀i Engine, developed by Cameroonian artist Joshua Kroon and Emmanuel Apetsi from Ghana, takes a more anthropological approach, working directly with the Baka community to preserve endangered polyphonic traditions using AI technology. Bina.ai, created by Nigerian strategist Ehinome Ogbeide and DRC technologist Muhigiri Ashuza Albin, positions itself as an AI-powered children’s music and storytelling platform grounded explicitly in African genres.
Two additional projects—Heritage in Code (developed by Kenyan deejay Linda Nyabundi and Ethiopian researcher Gebregziabihier Nigusie) and TIMah AI (led by Kenyan producer Tora Nyamosi and engineer Lawrence Moruye)—focus on preservation and archival, documenting African instrumental heritage while ensuring contributor royalties and community consent. These aren’t generic AI tools retrofitted for African use; they’re purpose-built systems that centre African voices, values, and ownership from the ground up.
As Apetsi observed during the presentations, “The future of music technology doesn’t have to be imported. It can be homegrown, collaborative and unmistakably African.” Prof Christo Doherty, who leads the initiative through the Wits Innovation Centre, reinforced this perspective, noting that whilst AI poses regulatory challenges, it also offers tremendous possibilities for African musicians if developed locally and contextually.
What gives these prototypes real significance is that they’re not merely technological experiments. They represent a deliberate effort to shift who gets to decide how music AI develops and, crucially, who benefits when it does. The projects currently remain in prototype phase with ongoing mentorship from the Wits institutions, but they signal a broader movement: African creators and technologists are determined to build the future of music technology themselves rather than accepting whatever solutions Silicon Valley and other international hubs dictate. For a continent whose musical influence shaped global culture for generations, claiming agency in the AI era isn’t optional—it’s essential.