How to adopt a multistakeholder approach to AI governance in Southern Africa?

Involving multiple stakeholders in the policymaking process creates a wider space for learning, deliberation, and the development of inclusive and informed decisions. It encourages policymakers to consider diverse perspectives, analysis and expertise, while preventing the polarization of policy discourse. The participation of multiple stakeholders coming from various backgrounds is essential for the development of relevant and efficient policies in each national context.

Involving all parties in the policy formulation process also ensures that it benefits everyone. Makaziwe Makamba emphasizes the importance of listening to the voices of underrepresented communities to bridge the digital divide and increase awareness of AI and its implementation. This recommendation is supported by the observation that across the world and particularly in the Global South, people – especially girls and women – are at risk of being excluded from the benefits of digital transformation. Inequalities reflected in AI systems frequently mirror the inequalities present in society. Hence, all stakeholders must be involved in the policy-making processes to ensure that their views and opinions are reflected in policy outcomes.

Elaborating AI policies with multiple stakeholders can be challenging. For example, definitions of AI may diverge from one party to another. UNESCO and i4Policy co-developed 10 building blocks to facilitate this process and bolster the foundations of inclusive policy design across the agenda setting, drafting and implementation of AI policies.

AI adoption must be supported, rooted in Southern Africa’s reality, and tailored for country and region-specific needs. For instance, entrepreneurs must consider local opportunities and challenges to understand what AI can deliver and what it cannot. Policymakers must integrate different perspectives to learn how AI can be misused and how its potential can be harnessed for national digital strategies and action plans.

Furthermore, context-relevant and evidence-based research must systematically inform AI governance and the development of applications in education, health and agriculture, among other domains.