AI, Livelihoods, and Building Futures: A Call to Action From Young African Women Scholars

Reflections from Kigali’s Deep Learning Indaba 2025 on Caribou’s commitment to facilitating inclusive AI scholarship.

These were Rashidat Ibrahim’s reflections at the end of Deep Learning Indaba, an annual gathering of the African machine learning and Artificial Intelligence (AI) community. Rashidat is one of the seven women PhD scholars in Caribou’s Scholars Collective in partnership with Mastercard Foundation. Her focus is on strengthening AI ethics in Nigerian higher education to ensure students don’t lose critical thinking and therefore employability in an AI age.

The Scholars Collective comprises women PhD students registered across African universities in Cameroon, Kenya, Nigeria, South Africa, and Tanzania. Their topics range from exploring generative AI chatbots in teaching mathematics to children on the autism spectrum disorder in Cameroon, to AI for disease detection in seaweed farming that may help women entrepreneurs in Zanzibar, Tanzania. Over the next year, the Caribou team will meet monthly with the Scholars and guest speakers in AI, working towards a collective publication on AI and Livelihoods.

While important AI interventions were being explored at the event—such as use of AI in malaria detection or building under-resourced language datasets—there was an urgent need to keep asking about the end goal of these and to move beyond proof-of-concept notes. Who will use these? How would we build skills? What are the barriers to overcome? 

AI needs to be seen as a system, not just a tool. For instance, AI as a system could shape the future of work and livelihoods for young Africans. Cynthia Amol, one of the scholars from Maseno University in Kenya, is writing her chapter on using generative AI to summarize under-resourced Kenyan language interfaces for government services young people may use in Kenya. Tatapong Belaya is looking at the tension between teaching aspiring creatives to use AI for script writing in Cameroon while ensuring “African” and Cameroonian identities are not lost. These use cases open up both the potential as well as the challenges of AI across the continent.

The focus in the Indaba community was on building the tech and datasets. This was apparent in the research poster sessions (standard sessions at academic conferences where attendees are invited to submit their research thesis summarized as posters through a selection process). Out of almost 300 posters, around 80% had a technical focus. This felt overwhelming for most of the scholars, who brought a social science perspective as to how the tech would impact the future of work and livelihoods. 

Yet the scholars held a lot of influence in new connections. For example, Sarah, working on AI and agricultural information, felt she had successfully challenged a physicist who was using AI to identify good thermoelectric materials, shifting their focus beyond merely building with AI to addressing the diverse needs of its end users, including cultural context and affordability. Myra connected with a Nigerian speech-to-text specialist who is developing a dataset for children with various disabilities, and they plan to explore opportunities for collaboration.

One of the key challenges in generative AI is the capture and replication of dominant forms and sources of knowledge. Carina Kanbi, a Ghanaian scholar, is working on indigenous knowledge in textiles and how it is captured and impacted by AI. One evening after the conference, the scholars visited Kimironko market in Kigali. In a meeting the following day, Carina highlighted how some Igbo words on a textile at the market were incorrect and outlined how this is something that could be reproduced by AI. There could be a serious cultural impact of incorrect data captured in datasets, which could come across as a “single truth.” 

This was related to a discussion on the scholars’ own use of AI. The seven scholars have been using AI to both write and correct their work, and the wrap-up conversation surfaced the drivers as to why (there was a shared view that it made things “sound better” or was more structured). However, the scholars also accepted that the risk could be losing seven individual perspectives and cultures if we relied on AI too much in our intellectual work. We agreed on an “Honesty Principle” when using AI, ensuring critical thinking in evaluating the AI output. 

In his book Superagency, Reid Hoffman makes the distinctions between “doomers” and “bloomers” of AI. These scholars take a more measured approach in between. Our aim for the Collective is to ensure that young scholars bridge and influence both academia and practice in the fast-moving, complex space of AI and livelihoods. In Myra’s words, “let the research not sit on the shelf.” 

Find out more about the Scholars Collective and explore our project on AI and the Future of Work.