In a groundbreaking fusion of technology and tradition, the University of the Witwatersrand (Wits) has unveiled the first prototypes from its AI & African Music (AIAM) project, marking a pivotal moment for higher education innovation in South Africa. Held on 16 April 2026 at the Chris Seabrooke Music Hall, the dual academic and industry showcases highlighted five artist-engineer teams' creations, blending artificial intelligence with diverse African musical heritages. This initiative, led by the Wits Innovation Centre (WIC) and the Machine Intelligence and Neural Discovery (MIND) Institute, underscores Wits' commitment to culturally informed AI development, addressing global biases in music generation tools dominated by Western datasets.
The AIAM project responds to the urgent need for Africa-led AI advancements in music, where current systems often struggle with complex polyphonies, rhythms, and languages from the continent. By fostering collaborations across seven African nations—South Africa, Cameroon, Democratic Republic of Congo, Ethiopia, Ghana, Kenya, and Nigeria—the prototypes not only preserve endangered traditions but also pioneer new creative frontiers. Professor Christo Doherty, holder of the Angela and David Fine Chair in Innovation, emphasized, "Artificial intelligence offers tremendous possibilities for African musicians... The Wits AI and African Music project created an opportunity for African musicians to work together with African AI engineers." Supported by US-based Wits alumnus Charles Goldstuck, the six-month pilot launched in November 2025 after selecting winners from over 150 applicants.
The Genesis of AIAM: Bridging AI and African Creativity
Wits University, one of South Africa's premier research institutions, has long championed interdisciplinary innovation. The AIAM project emerged from recognition that generative AI music tools like those from global tech giants falter on African sounds due to skewed training data. Sub-Saharan Africa's recorded music revenues reached US$120 million in 2025, up 15.2 percent year-on-year, with South Africa accounting for 78.1 percent, per the IFPI Global Music Report 2026. Yet, AI's Western-centric biases risk marginalizing these vibrant markets, prompting Wits to prioritize ethical, community-centered design.
The pilot funded multidisciplinary teams to explore four themes: archiving and preservation, creative practice and collaboration, tool development and localisation, and genre innovation. Selection involved technical reviews, feasibility checks, and interviews, ensuring regional diversity and feasibility. This higher education-led effort positions Wits at the forefront of responsible AI research, training the next generation of engineers and artists while contributing to South Africa's National AI Strategy.
Team 1: The Bɛ̀bɛ̀i Engine – Reviving Baka Polyphony
Joshua Kroon, a Cameroonian multidisciplinary artist known for cultural documentation through the FRED Movement, partnered with Ghanaian AI/ML engineer Emmanuel Apetsi, founder of SISU AI and leader in multilingual LLMs. Their prototype, The Bɛ̀bɛ̀i Engine, is an interactive AI instrument co-created with Cameroon's Baka community, focusing on endangered polyphonic vocal traditions, Liquindi water-drumming, and yodeling.
This tool enables real-time 'musical conversations' between humans and AI, allowing performers to improvise while the system responds in authentic Baka styles. Apetsi noted, "Technology begins to carry the rhythm, stories, and spirit of our people... We're shaping how the world will experience African creativity." By digitizing these oral traditions, the engine safeguards them against cultural erosion, offering educational and performance applications. In higher education contexts, it exemplifies how Wits supports cross-continental research, blending ethnography with machine learning.
Team 2: Heritage in Code – Archiving Instruments with Equity
Kenyan DJ and producer Linda Nyabundi, director of House of Abundance, collaborated with Ethiopian AI researcher Gebregziabihier Nigusie, expert in low-resource ML for cultural preservation. Heritage in Code is a secure digital archive where communities upload traditional instrument sounds, enriched with cultural context. An AI fusion engine then blends these with contemporary tracks, distributing royalties to contributors.
This addresses ownership gaps in AI datasets, ensuring African instrumental heritage—like mbiras or djembe variants—fuels innovation equitably. Nyabundi's work amplifies women in hip-hop, aligning with the project's diversity goals. For South African universities, such tools enhance musicology programs, providing datasets for AI training while promoting economic models for creators.
Team 3: ZAZI – The Musical Digital Twin
South African Afro-futurist Umlilo teamed with Gideon Gyimah, a Ghanaian voice-AI specialist who fine-tuned models for African languages. ZAZI acts as a 'musical digital twin,' inputting an artist's voice, lyrics, or ideas to generate African-inspired rhythms, melodies, harmonies, stories, and visuals in real-time.
Designed for experimentation, it preserves traditions while enabling genre-blending, like amapiano with ancient motifs. Gyimah's financial AI background ensures scalable, ethical deployment. This prototype highlights Wits' role in local talent development, potentially integrating into performing arts curricula across Gauteng institutions.
Team 4: Bina.ai – Nurturing Young Minds with African Rhythms
Nigerian music executive Ehinome Ogbeide, who propelled artists like Rema, worked with Congolese technologist Muhigiri Ashuza Albin on Bina.ai. This platform crafts children's music and stories rooted in African folktales, rhythms, and visuals, promoting learning through play.
Age-appropriate and multimodal, it counters Western edutainment dominance, fostering cultural identity from early ages. Ogbeide's branding expertise ensures engaging outputs. In South African higher ed, it inspires edtech research, aligning with decolonized curricula pushes at universities like UCT and UKZN.
Team 5: TIMah AI – Kikuyu Heritage Secured
Kenyan AI producer Tora Nyamosi and engineer Lawrence Moruye developed TIMah AI, a web archive for Kikuyu traditional music. It captures field recordings, lyrics, stories, and context, with AI transcription and community consent protocols.
Prioritizing sovereignty, users control access, preparing data for education. Moruye's speech tech focus tackles tonal complexities. This model influences SA's digital humanities, aiding preservation at institutions like Stellenbosch University.
The Showcase: A Catalyst for Dialogue
The 16 April events drew academics, students, and industry leaders. The academic session featured MIT's Anna Huang and a panel on AI ethics; industry included Goldstuck and execs from Boomplay, discussing sovereignty. Live demos and performances sparked debates on data bias—where African polyphonies confuse Western AIs—and royalties in generative tools. Explore the AIAM site for event recaps.
Ethical Imperatives: Sovereignty and Bias Mitigation
AI music tools exhibit biases, poorly handling African styles due to dataset imbalances. Studies show over 80 percent of producers reject AI songs over authenticity concerns. AIAM embeds ethics: consent, provenance, royalties. Goldstuck affirmed, "Africa... origin of so much early influences... enhance African human creativity." This aligns with SA's AI policy drafts, positioning Wits as a leader.
- Community co-creation prevents exploitation.
- Royalties ensure economic fairness.
- Local datasets counter Western dominance.
- Transparency in AI training fosters trust.
Implications for South African Higher Education
Wits' AIAM exemplifies research translation, training postgrads in ethical AI. It boosts interdisciplinary programs in digital humanities, music tech, and ML. Amid SA's youth unemployment, it creates pathways in creative tech. Universities like UJ and UWC could replicate, enhancing NIRF-like rankings.
Future Horizons: Scaling African AI Music Innovation
Post-pilot, prototypes evolve into open-source tools, with plans for expanded cohorts. Wits eyes pan-African hubs, integrating into curricula. As Africa's music market grows, AIAM positions SA unis as global innovators, blending beats and bytes for cultural renaissance.
For aspiring researchers, Wits offers vibrant opportunities in AI-music intersections, from PhDs to labs. This project not only preserves heritage but propels South Africa's higher ed into AI's ethical vanguard.
Photo by Planet Volumes on Unsplash
