The Need for International AI Governance: Insights from the UN’s Governing AI for Humanity Report, from an African Perspective

Introduction

For many African countries, the transformative potential of artificial intelligence is matched by serious concerns about exclusion, inequality and the imposition of standards developed elsewhere.

AI is frequently presented as a universal technology, yet the conditions under which it is designed, trained, deployed and governed remain deeply unequal. The benefits of AI are concentrated in jurisdictions with access to capital, research institutions, computing infrastructure, specialised skills and large technology markets. The risks, however, are distributed globally.

The United Nations’ Governing AI for Humanity: Final Report of the High-level Advisory Body on Artificial Intelligence offers an important global assessment of these challenges. It argues that AI’s cross-border nature, combined with its concentration in the hands of a relatively small number of States and corporations, requires forms of international cooperation that are more inclusive, coherent and accountable than those currently in place.

This research note examines the report’s principal findings and recommendations from an African perspective. It considers the international governance deficit, the limits of national and regional responses, and the risks that arise when African governments, experts, innovators and communities are excluded from decisions that will shape the continent’s technological future.

The report’s purpose and structure

The report was prepared by a high-level advisory body comprising experts from governments, academia, civil society and industry. Its central purpose is to identify gaps in existing approaches to AI governance and to propose practical mechanisms for addressing them.

The report focuses particularly on:

• the existing deficit in international AI governance;
• gaps in representation, coordination and implementation;
• emerging risks of exclusion and entrenched power imbalances;
• the relationship between AI governance, human rights and international law; and
• institutional proposals for developing more inclusive and effective international governance.

The report recognises that neither national regulation nor limited international groupings are sufficient to address the global reach of AI. AI systems, data, compute infrastructure, supply chains, cloud services and applications frequently operate across multiple jurisdictions. The effects of these systems may therefore extend well beyond the territory in which they were developed or deployed.

The report also draws attention to what may be described as an epistemic inequality in international standard-setting. Those with the greatest financial, technical and institutional resources are often best positioned to define the risks, priorities and governance models that others are subsequently expected to adopt.

For Africa, this is not an abstract concern. It raises the question of whose knowledge, experiences and interests are treated as authoritative when global AI standards are developed.

The international governance deficit and Africa’s exclusion

One of the report’s most significant findings is that international AI governance is currently shaped by a relatively small group of States and corporations. The report observes that at least 118 countries, many of them in Africa and the Global South, remain outside core international decision-making processes.

It further notes that seven States, namely Canada, France, Germany, Italy, Japan, the United Kingdom and the United States, participate in all of the key sampled AI governance initiatives. By contrast, every African State except one is excluded from these forums.

This exclusion is consequential. Decisions concerning AI policy, ethical standards, risk classification, technical safeguards and the future direction of the technology are increasingly made without meaningful participation from African governments, experts, civil society organisations and affected communities.

African priorities are consequently underrepresented in global AI debates. These priorities include:

• the relationship between critical minerals, natural-resource extraction and the AI supply chain;
• the extraction and use of African data;
• skills shortages and the migration of technical expertise;
• the underrepresentation and exploitation of African labour in AI development;
• data sovereignty and digital self-determination;
• the unequal distribution of AI infrastructure and compute capacity;
• algorithmic discrimination against people of African descent; and
• the effects of AI on marginalised, rural and indigenous communities.

If international AI governance continues to develop without meaningful African participation, the continent risks experiencing a renewed form of technological dependency. Data, labour, natural resources and markets may be incorporated into global AI systems without corresponding control, ownership or benefit-sharing.

This outcome is not inevitable. It is, however, a credible risk where governance structures reproduce existing asymmetries of wealth, knowledge and political power.

Why national efforts are necessary but not sufficient

Several African countries have adopted national AI strategies, established digital governance institutions or begun developing regulatory frameworks. The African Union has also advanced continental approaches to digital transformation, data governance and emerging technologies.

These efforts are essential. International governance cannot substitute for domestic accountability, regional cooperation or locally responsive regulation. African States must retain the capacity to govern AI in accordance with their own constitutional systems, public interests and development priorities.

At the same time, national and regional initiatives face significant limitations in responding to AI’s global infrastructure and cross-border effects.

First, AI infrastructure and supply chains are transnational. Data, compute capacity, semiconductor manufacturing, cloud services, technical platforms and research networks frequently lie outside the jurisdiction of the States in which AI systems are ultimately used. In Africa, much of this infrastructure remains located outside the continent, although African countries possess important assets in the form of data, human capacity, markets and natural resources.

Secondly, large technology and data corporations often possess resources and influence that exceed those of many States. Multinational companies may shape the development and use of AI systems in African markets without being subject to effective oversight by the governments in which those systems operate.

Thirdly, AI creates transboundary risks. Cybersecurity threats, privacy violations, discriminatory systems, disinformation, labour exploitation and harmful automated decision-making cannot be managed effectively by one State acting alone.

The appropriate conclusion is therefore not that local and regional governance should be deprioritised. It is that local, regional and international governance must operate together. Domestic and continental frameworks are necessary for accountability and contextual legitimacy, but an international layer is also required to address risks that exceed territorial borders.

The UN, despite its institutional limitations, remains an important strategic forum for developing that layer because of its universal membership and its formal commitment to human rights and international law.

Key risks for Africa in the absence of inclusive international governance

The report identifies several risks associated with an exclusive and fragmented approach to global AI governance. These risks have particular significance for Africa.

1. Data exploitation

African personal, economic, social and cultural data is increasingly valuable to global AI developers. Yet data may be collected, transferred and used without meaningful consent, transparency, ownership or benefit-sharing.

This creates the possibility of extractive data practices that resemble earlier forms of colonial exploitation. The underlying resource has changed, but the pattern may remain familiar: value is extracted from African societies, processed elsewhere and converted into commercial or strategic advantage outside the continent.

An increasing number of African States have adopted data protection legislation. However, the existence of legislation does not necessarily translate into effective protection. Regulators may lack the technical expertise, financial resources and jurisdictional reach required to monitor complex AI systems or hold multinational corporations accountable.

African States may also have limited leverage once data has been incorporated into systems that are developed, trained or operated elsewhere. Effective governance must therefore address not only data collection, but also downstream use, model development, cross-border transfers, automated inference and commercial benefit-sharing.

2. Widening digital and economic divides

AI is likely to accelerate economic opportunity in jurisdictions with access to research institutions, high-quality data, infrastructure, investment and specialised skills. Without deliberate intervention, African States risk being left behind in the emerging digital economy.

The danger is not only that Africa may become a passive consumer of AI products. African countries may also become sources of raw data, low-cost labour, natural resources and test markets while higher-value research, ownership and decision-making remain concentrated elsewhere.

International governance should therefore include measures that support African research capacity, infrastructure development, responsible innovation and participation in AI value chains. Development assistance should not be limited to technology transfer. It should also support institutional autonomy, local expertise and the ability to shape how technology is used.

3. Weakened digital autonomy

International standards and regulatory models developed without African participation may impose compliance costs and legal obligations that do not reflect the continent’s diverse legal, cultural, economic or institutional realities.

This creates a risk of normative misalignment. A framework may be formally universal while reflecting the assumptions of a limited number of jurisdictions. It may also prioritise risks that are highly visible in wealthy economies while overlooking risks that are more immediate in African contexts, including informal work, limited connectivity, institutional constraints, public-sector resource limitations and the use of AI in unequal social environments.

The result may be a reduction in digital sovereignty and digital self-determination. African States could find themselves implementing rules that they did not meaningfully help to develop, using institutions that they do not control and within markets where the principal economic benefits accrue elsewhere.

4. Labour, skills and resource extraction

Africa’s position in the global AI economy cannot be assessed only in terms of access to digital services. It must also be considered in relation to the labour and resources that sustain AI systems.

AI development depends on data labelling, content moderation, software development, technical maintenance, minerals and energy. African workers may perform essential forms of digital labour without receiving a proportionate share of the value generated. African countries may also supply critical minerals required for digital infrastructure while remaining marginal participants in the ownership and governance of the resulting technologies.

A credible international governance framework should therefore address labour rights, fair compensation, skills development, responsible supply chains and the environmental consequences of AI infrastructure.

Principles for international AI governance

The report anchors international AI governance in inclusion, human rights and international cooperation. It argues that AI governance should not be designed solely around the interests of technologically powerful States and corporations.

Its central premise is that AI should be governed in a manner that allows all States and communities to participate in, and benefit from, technological development.

The report places the UN at the centre of its proposed governance architecture. This is institutionally understandable, given the UN’s universal membership and its role in articulating international legal and human rights principles.

From an African perspective, however, this proposal must be approached with both strategic engagement and institutional realism. Although African countries represent the largest regional grouping within the UN, the continent remains underrepresented in many international decision-making structures.

The UN also carries a complicated historical legacy. It was established at a time when most African territories remained under colonial rule, and its institutional structures have often been criticised as reflecting unequal distributions of power. International law scholars have long examined the gap between the formal equality of States and the unequal conditions under which international rules are created and applied.[1] More recently, civil society organisations and commentators have criticised perceived double standards in the application of international law and humanitarian principles.[2]

These concerns do not make the UN irrelevant. They demonstrate why African participation must be substantive rather than symbolic. A UN-centred approach to AI governance will only be legitimate if African States and African non-State actors are able to influence its priorities, procedures and outcomes.

Mechanisms and recommendations with African relevance

The report proposes several mechanisms that could support a more coherent and inclusive approach to international AI governance.

1. A United Nations AI Office

The report recommends the creation of an agile AI office within the UN Secretariat. Such an office could coordinate AI-related work across the UN system, promote coherence between international initiatives and serve as an institutional link between governments, experts and other stakeholders.

For Africa, this office could provide a permanent institutional entry point for African priorities. Its effectiveness would depend on its mandate, resources and representation. It should not become another centralised body in which decisions are made primarily by technologically powerful States and corporations.

An AI office should therefore include meaningful African representation, transparent procedures, public reporting obligations and mechanisms for engagement with African regional institutions, national regulators, researchers and civil society.

2. Capacity development networks

The report places emphasis on building technical, regulatory and institutional capacity in regions that are currently underrepresented.

For Africa, capacity development should go beyond the transfer of technology or the delivery of short-term training programmes. It should support:

• African AI researchers and research institutions;
• regulators and public-sector officials;
• judicial and legal professionals;
• universities and technical training institutions;
• independent civil society organisations;
• local entrepreneurs and innovators; and
• regional and continental policy networks.

Capacity development should also be designed around retention. Training African experts without addressing the structural incentives that drive skills migration may simply strengthen institutions elsewhere.

The objective should be to build durable African expertise capable of shaping policy, developing technology, evaluating risk and participating in international standard-setting.

3. International data frameworks

The report calls for stronger international approaches to data governance. For Africa, this is particularly important because data should not be treated as an unlimited resource available for extraction without control or benefit-sharing.

International data frameworks should address:

• lawful and transparent data collection;
• meaningful consent and participation;
• cross-border data transfers;
• protection of vulnerable groups;
• indigenous and community data;
• access to data for public-interest research;
• accountability for downstream use; and
• equitable distribution of benefits generated from African data.

Such frameworks should also recognise that data governance is connected to development, trade, public administration and human rights. The objective should not be to isolate African data from international collaboration, but to ensure that collaboration takes place on fairer and more accountable terms.

4. Global scientific and policy panels

The report recommends structures that can provide independent scientific and policy advice on AI.

African scientists, policymakers, civil society organisations and affected communities should participate directly in these structures. Their participation should not be limited to consultation after the principal decisions have already been made.

African experts bring important perspectives on issues such as development, inequality, public-sector capacity, language diversity, informal economies, indigenous knowledge and the use of technology in conditions of limited infrastructure.

Global expertise is more credible when it reflects the diversity of the societies that AI systems affect. Representation should therefore be understood not only as a matter of fairness, but also as a condition of technical and regulatory quality.

From exclusion to action

As AI develops, international governance will become increasingly important to global safety, economic stability, human rights and social justice. The central issue is not whether international governance is necessary, but who will shape it and whose interests it will serve.

African States and institutions should therefore pursue a coordinated agenda that includes:

• strengthening African participation in international AI governance initiatives;
• developing common African positions on data, infrastructure and AI regulation;
• supporting the African Union’s role in global technology diplomacy;
• investing in local research, technical education and regulatory capacity;
• protecting African data and promoting fair benefit-sharing;
• addressing labour exploitation in AI supply chains;
• requiring transparency from multinational technology companies; and
• building coalitions with other countries in the Global South.

Africa has much to gain from AI. It also has significant knowledge, labour, data, markets and natural resources that will contribute to the development of the global AI economy.

The continent should not be positioned merely as a recipient of standards developed elsewhere. It should participate in defining those standards, challenge rules that are contextually inappropriate and advance governance models that reflect its own constitutional, social and developmental priorities.

The demand for a meaningful seat at the table is therefore not a request for inclusion as a matter of courtesy. It is a demand for participation in decisions that will shape Africa’s sovereignty, economies, institutions and human rights for decades to come.

Disclaimer

Africa is a continent of 54 States, each with distinct histories, cultures, political systems, legal frameworks and economic circumstances. The generalisations made in this article are intended to support broad analytical discussion and do not capture the diversity or specificities of individual African countries, communities or peoples.

References

1. United Nations, Governing AI for Humanity: The Final Report of the High-level Advisory Body on Artificial Intelligence (June 2024). View report
2. Franck Kuwonu, ‘Four African Countries at the Founding of the UN in San Francisco in 1945’ (2020) UN Africa Renewal.
3. James Thuo Gathii, ‘The Promise of International Law: A Third World View’ (2021) 36(3) American University International Law Review, Article 1. View article
4. CIVICUS, ‘Gaza: Double Standards Erode the Principles of International Humanitarian Law and Undermine its Credibility’. View source
5. Sorcha O’Callaghan, Ayesha Khan, Kathryn Nwajiaku-Dahou, Theo Tindall, Leen Fouad and Cecilia Milesi, ‘Humanitarian Hypocrisy, Double Standards and the Law in Gaza’ (8 November 2023) Expert Comment, ODI. View source