A People-Centric Approach to AI in Africa Demands More Participation... From The People

This essay is part of a collection of reflections from participants in the Participatory AI Research & Practice Symposium (PAIRS) that preceded the Paris AI Action Summit. Read more from the series here.

The deliberations around AI at the Paris AI Action Summit indicate to me that our battle is not in winning the AI arms race but in assuring that public policies are strong enough to protect the interest of the African people amidst an AI revolution that will influence our economies, politics, and lives. We need an endeavor that will demand participatory policy-making and tech governance.

The Paris AI Action Summit was attended by several global leaders from the EU, US, and Canada, and some representatives from Asia, while many African leaders seemed to have stayed away. An announcement made by Smart Africa on February 4 leading up to the Paris Summit indicates that African leaders are preparing for the continent's own iteration — The Global AI Summit on Africa. Taking place in Rwanda, the gathering will bring together political and industry leaders in AI as well as civil society organizations. Aiming to “pioneer new approaches and forge transformative partnerships to harness AI for Africa’s inclusive growth, and create a safe and secure digital future for all,” the agenda is motivated by an ambitious forecast that “by 2030, AI is estimated to add $2.9 trillion to the African economy.”

However, similar to the first iteration of the AI Safety Summit at Bletchley Park in the UK, the Global AI Summit on Africa is exclusive and can only be attended by invitation. It specifically targets “CEOs, C-suite executives, industry leaders and current and former government officials.”

To ensure the Global Summit on Africa successfully meets its objectives, African leaders and their allies can learn from some of the discussions that took place at the Participatory AI Research and Practice Symposium, which took place just before the Paris AI Action Summit: namely, that a participatory approach can help African countries shape AI policies in a manner that avoids perilous socio-economic pitfalls. Some of these pitfalls include an approach that singles out AI as a standalone sector to be invested in rather than a general-purpose technology that can catalyze the realization of existing development agendas. Another pitfall is the lack of regulatory protections to attract investments to the detriment of a vulnerable young workforce on the continent. Participatory AI also gives space for a critical voice that can endorse legitimate global partnerships that are favorable for Africans.

Although African political leaders have a tough task at hand to capitalize on the ongoing AI revolution to meet critical development goals, the focus should not be on AI alone. When viewed as a general-purpose technology, the African Union Continental Artificial Intelligence Strategy should explicitly be oriented as a policy instrument supporting the flagship projects under Agenda 2026. This means working closely with teams across these projects within the AU. For example, important investments such as funding for linguistic departments that can supply AI training data on African languages are missed with an AI-first approach.

I must note that the most impactful innovations in Africa, such as MPesa, were developed by people focusing on addressing developmental gaps, namely, financial inclusion. The objectives of the AU flagship projects and, essentially, African people stand to benefit more from an approach to innovation that is open and engenders participation.

Equally critical, not all participants have the public’s interest in mind when engaging with African policymakers. Big Tech investments in the continent's digital transformation, while welcomed, should not interfere with the introduction and enforcement of regulatory measures to protect Africans against exploitative and extractive practices. Specifically the young working population is often masqueraded as the most critical in the early stages of the AI supply chain.

There are also many offers to help African governments close connectivity gaps while at the same time not having the necessary guardrails to the mining of massive amounts of data from Africans. Or, the less spoken about, anti-competition practices by Big Tech and Multinational AI Labs who may want to position themselves as the fittest to offer the technologies to improve public service delivery in African countries. Such partnerships should be entered with the larger public in mind. I hope the data thematic area at The Global AI Summit on Africa will have a nuanced approach to data protection and regulation, which does not have the usual affinity towards economic gains and fewer risks.

Now, this does not mean that African leaders should not seek out partnerships. There are crucial partnerships such as with journalists/newsrooms that should be pursued to achieve trustworthy AI in African countries. My point is that these partnerships are not nurtured enough on the continent. Often, academic institutions, civil society, and a small fraction of the private sector on the continent have criticisms about the approaches, strategic partnerships, and collaborations pursued by government leaders that are not beneficial to the African people. These insights are discarded or paid less attention as they are deemed to hold back progress.

The “people-centric, development-oriented and inclusive approach” declared in the AU Continental AI Strategy will require hefty time and financial investments. This point stood out most for me as presenters under the Participation, Power, and Resistance track at the Research Symposium in Paris shared their work. They highlighted the importance of avenues for the public to contest the development of certain technologies or the approaches taken to solve problems communities face.

One such investment would be for the African Union and its member states to seriously factor in institutionalized mechanisms that allow resistance and contestation from the various partnerships. Avenues for this, such as the Pan African Peoples Parliament, are already available within the structures of the AU, however, are under-utilized due to a consultation model favored by the AU Commissioners. The latter does not allow adequate deliberation with a wide range of stakeholders but instead only includes the usual assemblage of AI consultants.

The so-called AI revolution is not similar to other revolutions preceding it. I think African leaders recognize this; however, they struggle to do away with the old ways of governance that were necessary to protect the sovereignty of a continent just emerging from years of divisive colonial practices. Sometimes that required only consulting a few trusted sources. In this new era pioneering new approaches that are beneficial for all will require that we forgo some of those protective habits and extend the doors open to the people themselves.