Centre for AI and Multidiscipline Solutions in Africa

Clear shared principles (not just vague statements) Agreement on a core set of actionable principles like safety, transparency, accountability, fairness, and human oversight that countries and organizations commit to applying. Success means these aren't abstract, but tied to real implementation guidance. Alignment across regions (especially Global South inclusion) A strong outcome would be meaningful participation and influence from regions like Africa, not just representation. That includes recognition of local priorities data sovereignty, infrastructure gaps, language inclusion, and economic impact so governance isn't dominated by a few powerful countries.

• Safe, secure and trustworthy AI
• Open-source software, open data and open AI models
• AI capacity-building
• Social, economic, ethical, cultural, linguistic and technical implications of AI

Safe AI ensures protection Capacity-building ensures participation Understanding implications ensures responsible and inclusive use If one is missing, governance becomes weak for example, you can't have trustworthy AI without capacity, and you can't govern AI properly without understanding its societal impact.

Data governance and sovereignty Who owns, controls, and benefits from data is still unresolved. Many countries risk becoming data providers while value is created elsewhere. This affects fairness, innovation, and national autonomy. Compute and infrastructure inequality Access to high performance computing and cloud resources is highly concentrated. Without addressing this, capacity building efforts remain limited and global AI development stays uneven. Concentration of power in AI ecosystems A small number of companies and countries dominate advanced AI. This raises concerns about competition, dependency, and the ability of smaller economies to shape rules or benefit fairly. Environmental and energy impact AI systems especially large models consume significant energy and water. Sustainability is often overlooked but is critical, particularly for regions already facing climate challenges.

A major gap is limited policy and regulatory readiness. While AI adoption is growing, clear national frameworks on data protection, accountability, and AI standards are still evolving. This creates uncertainty for innovators and exposes citizens to risks such as misuse of data and biased systems. Data governance and sovereignty remain critical concerns. Much of the data generated locally is stored or processed in the region, reducing local control and limiting economic value creation. This is compounded by infrastructure and compute inequality, where limited access to high-performance computing and reliable digital infrastructure constrains local AI development. There is also a capacity gap a shortage of skilled AI professionals, researchers, and informed policymakers. This slows down the ability to build, regulate, and scale AI responsibly. Linguistic and cultural underrepresentation for example, limited inclusion of Kiswahili and other African languages in AI systems reduces accessibility and relevance for local populations. Opportunities: These same gaps create space for strategic growth. Tanzania can leapfrog by embedding responsible AI principles early, rather than retrofitting regulation later. There is strong potential to develop locally relevant AI solutions in sectors like agriculture, healthcare, and education.

It can bridge policy fragmentation by aligning different national and regional approaches. Today, countries are developing AI regulations independently, which risks inconsistency and barriers to innovation. The Dialogue can promote interoperable standards and shared principles, making it easier to collaborate across borders. It can amplify underrepresented voices, particularly from the Global South. By ensuring meaningful participation not just attendance the Dialogue can integrate diverse perspectives on data governance, infrastructure gaps, and local priorities. This helps create governance frameworks that are globally relevant and equitable. It can translate principles into action by facilitating concrete commitments. This includes joint initiatives such as: Cross-border research collaborations Shared AI safety and evaluation frameworks Capacity-building partnerships (training, infrastructure, knowledge exchange) The Dialogue can act as a trust-building mechanism between governments, private sector, academia, and civil society. Open exchange reduces suspicion and encourages transparency, which is essential for addressing sensitive issues like data sharing and AI safety. It can support collective responses to global risks, such as misinformation, cybersecurity threats, and the governance of advanced AI systems that no single country can manage alone.

It can act as a convening bridge, bringing together these fragmented initiatives into a more coherent ecosystem, reducing duplication and aligning goals. It can translate high-level principles into actionable pathways, especially for countries that lack implementation capacity by offering toolkits, policy templates, and technical guidance. It can strengthen Global South participation, ensuring that frameworks like those from the African Union are not peripheral but central to global discussions. It can promote accountability and follow up, tracking commitments made across different initiatives and encouraging measurable progress.

Governments: Provide policy direction, share national strategies, and commit to interoperable regulatory approaches. They can also identify priority sectors (e.g., health, education) for pilot collaboration. Private sector: Contribute technical expertise, safety practices, and resources. Companies can support standards development, transparency tools, and infrastructure partnerships. Academia and research institutions: Offer evidence based insights, independent evaluations, and foresight on emerging risks and technologies. Civil society and communities: Ensure human rights, inclusion, and local realities are reflected—especially for marginalized groups and underrepresented regions. International and regional organizations: Facilitate coordination, align existing frameworks, and support capacity-building across countries.

Global South countries, especially from Africa and small island states, whose priorities (infrastructure gaps, data sovereignty, development needs) differ from those of major AI powers. Local communities and grassroots organizations who experience AI impacts directly but rarely shape policy. Indigenous groups and cultural communities, whose knowledge systems and values are often excluded from AI design and governance. Youth and students, despite being the largest future users and workforce affected by AI. Women and marginalized groups, who are disproportionately affected by bias and digital exclusion. Small and medium enterprises (SMEs) and local innovators, who lack the resources to engage in global forums dominated by large tech companies. Non-English and low-resource language communities, whose linguistic needs are often ignored in AI systems. How to include them: Funded participation and access: Provide travel support, stipends, and virtual access to remove financial and geographic barriers. Regional consultations: Hold pre dialogues at regional and national levels to gather input that feeds into global discussions. Language inclusion: Offer translation, interpretation, and multilingual materials to enable meaningful participation. Structured representation: Allocate dedicated seats or quotas for underrepresented groups in panels and decision-making spaces.

Multi stakeholder "policy labs" Small, mixed groups of governments, industry, academia, and civil society work on real AI governance challenges (e.g., regulating deepfakes or AI in education). Each lab produces a concrete policy prototype or recommendation. Scenario simulation exercises Participants respond to future AI scenarios (e.g., AI-driven misinformation during elections or autonomous systems failure). This helps stress-test governance systems and build shared understanding of risks. Regional solution showcases Dedicated sessions where countries and local innovators present AI solutions from their contexts, especially from the Global South. This shifts the narrative from "knowledge recipients" to "solution providers." Interactive "commitment marketplaces" Stakeholders publicly present what they can offer (funding, data, tools, training, policy support) and what they need, encouraging direct partnerships and matchmaking. Live drafting sessions Instead of finalizing documents behind closed doors, key outputs (principles, frameworks) are co-drafted in real time with participant input, increasing transparency and ownership. Digital participation platforms

A key policy example is the UNESCO Recommendation on the Ethics of Artificial Intelligence, which provides a global normative framework emphasizing human rights, transparency, fairness, and accountability. Similarly, the OECD AI Principles promote trustworthy AI and have influenced many national AI strategies. On the regulatory side, the European Union's European Union Artificial Intelligence Act introduces a risk-based approach, categorizing AI systems by levels of risk and setting strict requirements for high-risk applications. This is one of the most structured attempts at binding AI regulation. For collaborative governance, the Global Partnership on AI brings together governments, industry, and researchers to develop practical guidance on responsible AI development and deployment. Similarly, the Internet Governance Forum provides an open platform for dialogue on digital policy, including AI. In terms of technical and safety approaches, AI auditing frameworks, model evaluation benchmarks, and "responsible AI toolkits" developed by companies and research institutions help identify bias, improve transparency, and assess system risks before deployment. Capacity-building platforms such as AI training hubs, open-source machine learning communities, and regional innovation centers also play a critical role in democratizing access to AI knowledge and tools, especially in developing regions.