AI AND THE QUESTION OF POWER: GHANA’S SEARCH FOR A GOVERNANCE POSITION

 INTRODUCTION

In a previous article on the global race for AI governance, I argued that artificial intelligence is not merely a technological contest, but a struggle over power — over whose values, interests, and philosophies shape the systems that increasingly govern economies and societies. Africa has yet to clearly articulate its own governance philosophy, and that absence is not neutrality. It is vulnerability.

In a subsequent analysis of continental efforts, particularly the Smart Africa AI Blueprint and the African Union's AI Strategy, I suggested that Africa is not without direction. An approach is emerging: development-driven, adoption-led, and capacity-focused. But it remains incomplete. It is clear on the uses of artificial intelligence, but less explicit on how it is to be governed as an instrument of power, and on whose terms that governance will rest. The question, then, is how this emerging approach translates at the national level. Ghana provides a particularly instructive case.

As one of the early movers in articulating a national AI strategy in Africa, Ghana offers a concrete expression of how these continental priorities are being interpreted and operationalised. Its National Artificial Intelligence Strategy (2025–2035), building on an earlier 2023–2033 framework, reflects a significant refinement of approach, with greater economic clarity, stronger institutional grounding, and a more structured pathway for implementation. The strategy retains its ambition to position Ghana as a trailblazer for AI leadership in Africa and beyond, and outlines a comprehensive programme for building the capabilities, infrastructure, and ecosystems required to support that ambition. But ambition, even when more fully articulated, does not define the form that leadership is intended to take. It is precisely that gap between stated ambition and defined position that this article examines.

This article does not seek to evaluate Ghana's strategy in terms of its technical merits, nor to propose an alternative framework. Rather, it uses Ghana as a lens through which to examine a broader question: what does Africa's emerging approach to artificial intelligence look like in practice, and what does it leave unresolved? It argues that Ghana's strategy, like the continental frameworks that inform it, is clear on purpose but less defined in its positioning, and that closing the gap between the two is not a technical task but a governance one. In doing so, it asks not only what Ghana intends to do with AI, but how it is positioning itself in relation to the systems, actors, and power structures that will shape the technology's future.

WHAT GHANA GETS RIGHT

Ghana's National Artificial Intelligence Strategy (2025–2035) demonstrates several important strengths that provide a solid foundation for its engagement with artificial intelligence. Building on an earlier 2023–2033 framework, the updated strategy reflects a process of refinement, with greater structural clarity and a more focused approach to implementation.

First, it establishes a clear development orientation. Artificial intelligence is positioned as a tool for economic transformation, improved public service delivery, and social inclusion. This grounding is significant. It ensures that AI policy is not treated as an abstract technological pursuit, but is instead aligned with Ghana's broader development priorities. This intent is captured in the strategy's mission:

"To harness AI for inclusive growth across all sectors and to improve the lives of people in Ghana, becoming a trailblazer for AI leadership in Africa and beyond."

Second, the strategy is structured and internally coherent. Defined pillars organise action across key areas such as infrastructure, data systems, skills development, innovation, and sectoral application. This structure enhances clarity and makes the strategy accessible to a wide range of stakeholders.

Third, there is a strong emphasis on capability and capacity building. The focus on digital infrastructure, data availability, and human capital reflects an understanding that meaningful AI adoption depends on foundational readiness. The attention given to skills development and research ecosystems suggests a recognition that long-term value lies not only in using AI, but in developing the capacity to engage with it more deeply over time.

Fourth, the strategy is grounded in practical application. By linking artificial intelligence to sectors such as agriculture, healthcare, education, and finance, it identifies areas where AI can deliver tangible and near-term impact. This focus strengthens the strategy's relevance and increases the likelihood of measurable outcomes.

The updated strategy also reflects a shift toward greater economic and institutional specificity. Artificial intelligence is framed not only as an enabler of transformation, but as a measurable driver of growth, with explicit emphasis on investment mobilisation, ecosystem expansion, and long-term economic contribution. This introduces a clearer linkage between AI policy and national economic strategy, complemented by strengthened coordination mechanisms and institutional structures that signal a move from policy articulation toward sustained implementation at scale.

Taken together, these elements point to a strategy that is purposeful, structured, and increasingly operational in its design. This is a foundation that deserves recognition. At the same time, the mission introduces a broader ambition, one that extends beyond application to a claim of leadership. It is that transition, from development purpose to leadership ambition, that requires closer examination.

WHAT THE STRATEGY SIGNALS (IMPLIED PHILOSOPHY)

What is at issue here is not the absence of policy direction, but the absence of a clearly articulated governance philosophy. In this context, a governance philosophy does not refer simply to regulatory frameworks, ethical guidelines, or implementation mechanisms. Rather, it reflects a coherent position on how a country understands and exercises authority, control, and influence over artificial intelligence systems, including how it approaches questions of ownership, dependency, participation, and power within the global AI ecosystem. It is this level of articulation, the translation of intent into a defined position, that remains less clearly expressed in Ghana's strategy.

The structure and priorities of the strategy nevertheless point to a clear underlying approach, one that is consistent with, and in many respects a direct reflection of, the continental framework already identified in the African Union's AI Strategy. It is one that is development-first, adoption-led, capacity-driven, and partnership-dependent.

It is development-first. Artificial intelligence is consistently framed as a tool for economic transformation, improved public service delivery, and social inclusion. The emphasis is not on advancing the technological frontier, but on applying AI to address practical challenges within the Ghanaian economy. This aligns AI policy closely with national development priorities.

It is adoption-led. The strategy prioritises the deployment of existing technologies across sectors rather than the development of frontier AI systems. Its focus is on integration, application, and scaling. In this respect, Ghana is positioned primarily as a user and implementer of AI, rather than a producer of core technologies.

It is capacity-driven. Significant attention is given to building the foundational elements required for AI adoption, including digital infrastructure, data systems, skills development, and research ecosystems. This reflects an understanding that sustained engagement with artificial intelligence depends on strengthening domestic capabilities over time.

It is partnership-dependent. The strategy places strong emphasis on collaboration with international partners, private sector actors, and global technology providers. This reflects both a pragmatic recognition of current capacity constraints and an openness to integrating into existing global AI ecosystems.

This combination reflects a deliberate and pragmatic approach to engaging with artificial intelligence within existing constraints, and it is closely aligned with the broader continental pattern already identified in previous analyses. But it is, by design, an approach centred on use rather than control. What remains less clearly articulated is how this approach translates into a governance position, particularly in relation to sovereignty, influence, and the terms of participation within the global AI ecosystem. That is the gap this article examines.

CLARIFYING THE MEANING OF LEADERSHIP

This question of leadership sits at the centre of the strategy's stated ambition. The mission's reference to Ghana as "a trailblazer for AI leadership in Africa and beyond" invites closer scrutiny: leadership in what sense?

In the context of artificial intelligence, leadership is not a generic aspiration. It has distinct and identifiable forms. A country may lead in the development of frontier technologies, in the deployment of AI across key sectors, in the shaping of regulatory standards, or in defining how the technology is governed within its economy and beyond. Each of these reflects a different position within the global AI ecosystem, with distinct implications for capability, control, and long-term influence.

The strategy implicitly aligns Ghana most closely with leadership through deployment and application. The emphasis on skills, infrastructure, data systems, and ecosystem development reflects a coherent effort to integrate AI into the national economy, grounded, internally consistent, and clear on what AI is for and how it can be used to support development. But the strategy does not explicitly define whether this is the intended strategic position or a transitional stage toward a broader form of influence. That distinction matters, and it remains unresolved.

But leadership implies more than effective adoption. What remains less clearly articulated is whether Ghana intends to shape the conditions under which AI operates, including the rules that govern its use, the standards that define its deployment, and the terms of participation for external actors in the domestic digital economy. These are not questions of application. They are questions of position and control.

The absence of this distinction does not invalidate the strategy. It reflects a broader pattern already visible at the continental level, one in which ambition is clearly expressed, but its form remains less defined. Without that clarity, "leadership" risks remaining an aspirational label rather than a strategic position, and aspirational labels, however sincerely held, do not shape the terms on which technology is governed. Defining what leadership means, in practice and in position, is the work that remains.

WHAT IT LEAVES OPEN

The gaps in Ghana's strategy are not isolated technical omissions. They are expressions of the same unresolved strategic choices that confront Africa at the continental level, choices that, when left implicit, shape governance by default rather than by design. Examined through that lens, five areas stand out.

·       Sovereignty or Dependency

The strategy recognises the importance of data, infrastructure, and digital systems (including the expansion of national datasets, cloud and compute capacity, and local language data resources). But it does not clearly articulate who ultimately controls these assets, or how that control is to be exercised. As AI systems increasingly rely on data flows and computational infrastructure that extend beyond national boundaries, the question of ownership and authority becomes central. Without a defined position, sovereignty remains acknowledged but not operationalised.

·       Enablement or Control

The strategy addresses risk, ethics, and coordination through the proposed Responsible AI Authority and alignment with the UNESCO AI Ethics Recommendation. But it is less explicit on how authority is to be exercised over the systems being adopted, who defines the rules, who enforces them, and how conflicts between domestic priorities and external interests are to be resolved. The governance framework is oriented toward coordination and ethical oversight rather than toward the exercise of regulatory control as an instrument of power. These are not technical questions. They are questions of position.

·       State or Market Leadership

The strategy adopts a multi-stakeholder, partnership-dependent model, drawing on government, private sector, academia, and international partners through public-private partnership arrangements. This reflects a pragmatic and necessary approach given existing capacity constraints. However, it leaves less clearly defined the terms on which these partnerships are structured. Who sets the conditions of engagement? How are national interests protected within externally driven systems? The balance between state authority and market dynamics is present in the strategy's design but not resolved as a deliberate governance choice.

·       Fragmentation or Coordination

The strategy is designed for Ghana as a national framework. It references global collaboration and international partnerships, and its governance model is inspired by approaches in the UK, Singapore, and Egypt. But it does not define Ghana's position within a coordinated continental governance architecture. The risk is not internal fragmentation, but external alignment, a nationally coherent strategy that integrates into global ecosystems without a clearly defined continental anchor. Participation in global AI ecosystems does not, in itself, guarantee influence.

·       Adoption or Influence

By prioritising adoption and integration across key sectors, agriculture, healthcare, finance, and public administration, the strategy positions Ghana primarily as a user and implementer of AI rather than a shaper of the conditions under which AI is governed. The strategy signals an emerging shift toward producer-led innovation through the NLP Centre of Excellence and the National Deep Science Institute. But it does not fully define whether Ghana intends to shape the standards, norms, and commercial arrangements that govern those ecosystems. Participation, in itself, does not guarantee influence.

Taken together, these are not gaps in implementation. They are areas where the strategy remains open, whether by design or by omission. They define the distance between a strategy that enables the use of AI and a governance philosophy that shapes the terms of that use. That distance is what this article examines.

FROM INTENT TO ARTICULATION: A GAP IN TRANSLATION

The absence of a clearly defined governance position within the strategy does not necessarily indicate the absence of a broader philosophical orientation. In fact, elements of such a philosophy are visible in the political articulation that accompanies the strategy.

In his address at the launch of the National Artificial Intelligence Strategy (2025–2035), President Mahama emphasised that Ghana must not remain a passive consumer of externally developed technologies, but must actively "build, own, and govern" artificial intelligence systems that reflect its values, languages, and developmental priorities. This framing points toward a governance philosophy that is more explicitly concerned with sovereignty, control, and localisation than is consistently reflected in the strategy itself.

This distinction is significant. The strategy, as structured, emphasises development, adoption, capacity, and partnership. It focuses on enabling the use of artificial intelligence within the national economy and building the systems required to support that use. In contrast, the political articulation places greater emphasis on ownership, autonomy, and the terms of participation in the global AI ecosystem. The consequence of that contrast is not merely rhetorical; it affects how the strategy is read, how its priorities are weighted, and how implementation decisions are made when external pressures and domestic intent pull in different directions.

The result is not a contradiction, but a gap in translation, between articulated intent and institutionalised position. The underlying philosophical intent, to move beyond passive adoption toward a more controlled and locally grounded AI ecosystem, is present, but not fully codified within the strategic framework. As a consequence, the strategy appears more cautious and integration-oriented than the broader ambition it is intended to support.

This gap is consequential. Governance philosophies shape how policies are interpreted, prioritised, and implemented over time. Where such philosophies remain implicit, execution risks defaulting to existing global patterns rather than reflecting the distinct position that political leadership seeks to establish. Closing that gap, by translating intent into a codified position, is not a technical exercise. It is the governance work that gives ambition strategic meaning.

THE RISK

What emerges from this is not a failure of strategy, but a limitation of scope. An approach that is development-oriented, adoption-led, and partnership-dependent can deliver meaningful gains in the short to medium term. It can accelerate access to technology, enable experimentation, and support the integration of artificial intelligence into key sectors of the economy. In this respect, the strategy is well aligned with Ghana's immediate priorities. But over time, the absence of a clearly defined governance position introduces a different set of risks.

These are not risks that arise from what external actors may do to Ghana. They arise from what Ghana has not yet defined for itself. Structural possibilities that emerge when governance is not fully resolved at the level of position.

The first is the risk of structural dependency. Where core technologies, data infrastructures, and technical standards are shaped externally, domestic adoption may proceed within systems over which Ghana has limited influence. This does not prevent progress, but it can constrain the extent to which that progress is directed on nationally defined terms.

The second is the risk of externally shaped outcomes. In the absence of clearly articulated positions on control, standards, and participation, the evolution of AI systems within the domestic economy may be influenced more by external commercial and technological priorities than by internal policy intent. Over time, this can affect how value is created, distributed, and retained.

The third is the risk of participation without influence. The strategy positions Ghana within global AI ecosystems, but does not fully define how it intends to shape them. Engagement, in this sense, becomes reactive rather than strategic. It enables access, but does not necessarily confer leverage.

These risks are not immediate or inevitable. But they are structural, and structural risks compound over time. They do not negate the value of the current approach. They define its limits and the work that lies beyond them, work that begins with translating Ghana's stated ambition into a governance position that is as clearly defined as the development purpose it is intended to serve.

 

 

CONCLUSION

Ghana's engagement with artificial intelligence reflects a broader continental shift already underway. The evolution of its national strategy, particularly in its 2025–2035 iteration, demonstrates a deepening focus on development, capacity, and application. This reflects a clear recognition of the opportunities AI presents and the practical steps required to harness them. In this respect, the strategy is coherent, grounded, and aligned with national priorities.

But clarity of purpose does not, on its own, resolve the question of position. What remains open is not what artificial intelligence is for, but how it is to be governed, and on whose terms that governance will rest. This question is not only a matter of policy design, but of how underlying intent is translated into formal strategy. While elements of a governance philosophy are increasingly visible in broader political articulation, they are not yet fully or consistently codified within the strategic framework. The distinction is not technical. It is strategic. It is the difference between participating in systems shaped by others and shaping the conditions of participation itself.

There is no single template for AI governance. Even among the major global actors, approaches differ significantly. China's model reflects a more state-centric approach to control, the European Union emphasises rights-based regulation, the United States prioritises market-led innovation, and the United Kingdom adopts a more flexible, pro-innovation posture. Each engages with common principles, addressing issues of risk, safety, accountability, and innovation, but with different emphases. These differences are not accidental. They reflect underlying governance philosophies about how power is exercised within the digital economy.

What remains less clearly articulated is how Africa, and Ghana in particular, intends to position itself within this landscape. The elements of an approach are emerging, visible in the strategy's development orientation, in the institutional architecture being built, and in the political articulation that accompanies it. But an emerging approach is not yet a defined position. A position requires clarity on sovereignty, on the terms of partnership, on the balance between state and market authority, and on what influence, rather than merely participation, is intended to look like. Until those questions are answered explicitly, Ghana's governance philosophy will remain implied rather than institutionalised, present in intent, but not yet embedded in the frameworks that give intent its force.

The implication is not that Ghana must identify a "correct" model to follow. It is that governance cannot be assembled through imitation alone. Shared principles must be interpreted and prioritised through a clearly defined position, one that reflects domestic priorities, institutional realities, history, culture, and strategic intent.

Ghana has articulated ambition. It has begun to signal the contours of a governance philosophy. The question now is whether it will fully define, codify, and embed the position that gives that ambition strategic meaning. Because an AI strategy without a governance philosophy is like a blueprint without a foundation: it can describe what is to be built, but it cannot determine on whose ground, by whose rules, and for whose benefit that building will stand. The meal that will govern the world's AI future is already being prepared. Does Ghana want to be part of consuming it, or part of preparing it? That choice can change the policy, institutional, and execution dynamics.