AI4D Knowledge Synthesis

ABOUT THE SERIES ​ — Gender Knowledge Synthesis

Since its launch in August 2023, in collaboration with the AI4D Network, the exceptional Gender Knowledge Synthesis webinar series has focused on sharing valuable insights from African experts and researchers working on Artificial Intelligence and centering feminist methodologies in their practice and praxis. 

How can communities, social Scientists and technologists leverage AI and Human Rights-based approaches?  How can the African Union’s Maputo Protocol be leveraged to help create the environment for a more human rights based AI? This conversation dives deeper into how human Rights frameworks can be the starting point for the creation of new tech & innovation.

The Maputo Protocol is now 20 years old. Through its visionary 32 articles, it encompasses significant demands for the rights of women in Africa. By revisiting its four dual approaches: the elimination of harmful practices, the right to participation in decision making processes, education and training, and positive cultural contexts – and bringing them to the heart of the contemporary digital context – projects may be able to guide today’s tech to be free from bias and discrimination. 

With Angella Ndaka (Centre for African Epistemic Justice); Winston Ojenge (African Centre for Technology Studies, ACTS,Kenya) and Mitchel Ondili (Gender At Work & Women At The Table) | Moderator: Amina Soulimani 

Putting the needs of the community as a priority when designing AI tools is transformative. This panel brings together the outcome of six interventions from the AI4D Gender and Inclusive AI Research and Innovation Challenge innovation challenge, which work on the nexus of gender and AI in education, Intimate Partner Violence, agriculture, and healthcare on the African continent. In each of the projects,valuable methodological approaches have been deployed such as; context-based, project-based study,co-creation approach, and user-led inclusive design. The speakers explain ways to carry out feminist and gender responsive interventions, in every step of the development process, to reach inclusive outcomes.

With Gloriana Monko (Dodoma Lab, University of Dodoma); Winston Ojenge (African Center for Technology Studies | ACTS, Kenya); Eunice Akyereko Adjei (Responsible Artificial Intelligence Lab); Daisy Salifu (International Centre of Insect Physiology and Ecology | AI4AFS hub); Adekemi Omotubora (EDU AI); Wambui Gachiengo (Villgro Africa) | Moderator: Khanysa Mabyeka

This conversation, stirred and driven by three experts in education and economics, puts together a roadmap for inclusive interventions and advocacy plans that entail holistic learning environments, role models, and stress on communities’s critical role in creating broader participation opportunities. Most importantly, they highlight the importance of shifting the question of participation around to: how can STEM be reimagined to fit the ambitions of girls to access new labor markets?

With Adedeji Adeniran (Gender Responsive AI Network & Centre for the Study of the Economies of Africa); Alice Amegah (World Bank Fellow) and Gloriana Monko (Dodoma Lab, University of Dodoma)

How can we learn through feminist practice? Peer to peer learning is a method that challenges hierarchical modes of acquiring knowledge, knowledge production and dissemination. Often, gender training is not enough to witness transformational change.This conversation demonstrates the ways in which gender training, intersectionality and inclusion needs to be undertaken through methodologies that merge experience and theory.  4 experiences and points of view on experiences working through peer-learning to develop effective and gender responsive AI tools on the African continent. 

With Marie-Katherine Waller (Gender At Work); Daisy Salifu (AI for AFS); Kemi Omotubora (EDUAI); Caesar Wisdom Favor (HASH Lab) 

How can AI help women and men in the field of agriculture? Caroline Mbaya and Winston Ojenge from the African Center for Technology Studies (ACTS) in Kenya highlight the critical importance of rethinking the adequacy of technological tools developed for the farming industry and the diverse communities/users. For instance, linguistic diversity must be taken into consideration to avoid social exclusion or access to substantial information. Evidence-based research on the challenges that hinder the full participation of women and youth must be taken into account when developing AI solutions.

This conversation raises an important question: how well do AI developers/ entrepreneurs know about women and youth’s challenges pertaining to unpaid labor and care work? Caroline and Winston argue that there must be no double standards when developing farming solutions or when gathering the necessary data since remote or second-hand data collection can often not be fully reliable. 

In his dynamic discourse, Dr; Fuh calls for a renaissance in AI development, one that embraces feminist and decolonial thinking with the rich, diverse tapestries of African wisdom and perspectives. He speaks of “espitemic disobedience” urging us to consciously uncouple from the colonial legacies that have long skewed our understanding of knowledge itself.

Dr. Fuh’s discussion is a vibrant tapestry of ideas, weaving together the threads of human dignity, the challenge to systemic inequality, and the quest for a world where discrimination has no place. It’s a narrative that underscores the evolving journey of decolonial thought, a process that’s not static but dynamic, inviting us to question the established norms and embrace a plurality of voices and perspectives. This is no mere academic debate; it’s a clarion call for action. It’s about reimagining the very essence of AI in an African context, breaking away from the vestiges of the past to forge a path toward a future where technology is not just innovative but also inclusive and just. It’s about sustained efforts, a relentless drive to dismantle the biases that lurk within algorithms, and a commitment to a world where AI champions fairness and equity. Dr. Fuh’s vision is clear: a future where AI serves all of humanity, unshackled from the chains of its colonial and patriarchal legacies.

With Divine Fuh (University of Cape Town)

In this scintillating discussion, Ernest Mwebaze and Nyalleng Moorosi outline an optimistic future for Africa’s progress in addressing AI bias, data governance and related challenges in Africa. Ernest emphasises a need for internal, community based, community led initiatives, which will respond to externalised pressures coming from large companies, AI institutions, universities, researchers, and industrial labs outside Africa. 

There is a recognition that the allure of solving trendy problems brought by external influences can divert focus. The speaker expresses a perspective that values the ongoing internal examination of problems, even if they may seem relatively small in scale. The fear is that neglecting internal issues might lead to another situation where knowledge has to be imported, akin to a new educational revolution. The discussion highlights the gap in data protection laws across African countries and the need for policies with enforcement mechanisms. The tension between the global knowledge landscape and the need for local governance is explored, with a call for homegrown solutions.

With Ernest Mwebaze (AI4D Gender Responsive AI Network) and Nyalleng Moorosi (The Distributed AI Research Institute)

In this illuminating discussion moderated by Nyalleng Moorosi, in conversation with Kathleen Siminyu and Nombuyiselo ‘Mbuyi’ Zondi, the panelists bring into focus the unseen biases that occur when representation is not adequately considered. They expound on the need for an intersectional lens and a deconstruction of the notion that technologists solely hold the knowledge to develop models. Through an understanding of the diversity of language and dialects, they propose interdisciplinary collaboration to develop the ecosystem underlying different algorithms.
 
With Nyalleng Moorosi (DAIR), Kathleen Siminyu (Mozilla) and Nombuyiselo Zondi (University of Pretoria)