Artificial intelligence (AI) is increasing productivity and pushing the boundaries of what's possible. It powers self-driving cars, social media feeds, fraud detection and medical diagnoses. Touted as a game changer, it is projected to add nearly US$15.7 trillion to the global economy by the end of the decade.

Africa is positioned to use this technology in several sectors. In Ghana, Kenya and South Africa, AI-led digital tools in use include drones for farm management, X-ray screening for tuberculosis diagnosis, and real-time tracking systems for packages and shipments. All these are helping to fill gaps in accessibility, efficiency and decision-making.

However, it also introduces risks. These include biased algorithms, resource and labour exploitation, and e-waste disposal. The lack of a robust regulatory framework in many parts of the continent increases these challenges, leaving vulnerable populations exposed to exploitation. Limited public awareness and infrastructure further complicate the continent's ability to harness AI responsibly.

What are African countries doing about it? To answer this, my research mapped out what Ghana and Rwanda had in place as AI policies and investigated how these policies were developed. I looked for shared principles and differences in approach to governance and implementation.

The research shows that AI policy development is not a neutral or technical process but a profoundly political one. Power dynamics, institutional interests and competing visions of technological futures shape AI regulation.

I conclude from my findings that AI's potential to bring great change in Africa is undeniable. But its benefits are not automatic. Rwanda and Ghana show that effective policy-making requires balancing innovation with equity, global standards with local needs, and state oversight with public trust.

The question is not whether Africa can harness AI, but how and on whose terms.

Rwanda's National AI Policy emerged from consultations with local and global actors. These included the Ministry of ICT and Innovation, the Rwandan Space Agency, and NGOs like the Future Society, and the GIZ FAIR Forward. The resulting policy framework is in line with Rwanda's goals for digital transformation, economic diversification and social development. It includes international best practices such as ethical AI, data protection, and inclusive AI adoption.

Ghana's Ministry of Communication, Digital Technology and Innovations conducted multi-stakeholder workshops to develop a national strategy for digital transformation and innovation. Start-ups, academics, telecom companies and public-sector institutions came together and the result is Ghana's National Artificial Intelligence Strategy 2023-2033.

Both countries have set up or plan to set up Responsible AI offices. This aligns with global best practices for ethical AI. Rwanda focuses on local capacity building and data sovereignty. This reflects the country's post-genocide emphasis on national control and social cohesion. Similarly, Ghana's proposed office focuses on accountability, though its structure is still under legislative review.

Ghana and Rwanda have adopted globally recognised ethical principles like privacy protection, bias mitigation and human rights safeguards. Rwanda's policy reflects Unesco's AI ethics recommendations and Ghana emphasises "trustworthy AI".

Both policies frame AI as a way to reach the UN's Sustainable Development Goals. Rwanda's policy targets applications in healthcare, agriculture, poverty reduction and rural service delivery. Similarly, Ghana's strategy highlights the potential to advance economic growth, environmental sustainability and inclusive digital transformation.

Rwanda's policy ties data control to national security. This is rooted in its traumatic history of identity-based violence. Ghana, by contrast, frames AI as a tool for attracting foreign investment rather than a safeguard against state fragility.

The policies also differ in how they manage foreign influence. Rwanda has a "defensive" stance towards global tech powers; Ghana's is "accommodative". Rwanda works with partners that allow it to follow its own policy. Ghana, on the other hand, embraces partnerships, viewing them as the start of innovation.

While Rwanda's approach is targeted and problem-solving, Ghana's strategy is expansive, aiming for large-scale modernisation and private-sector growth. Through state-led efforts, Rwanda focuses on using AI to solve immediate challenges such as rural healthcare access and food security. In contrast, Ghana looks at using AI more widely - in finance, transport, education and governance - to become a regional tech hub.

The effectiveness of these AI policies is held back by broader systemic challenges. The US and China dominate in setting global standards, so local priorities get sidelined. For example, while Rwanda and Ghana advocate for ethical AI, it's hard for them to hold multinational corporations accountable for breaches.

Energy shortages further complicate large-scale AI adoption. Training models require reliable electricity - a scarce resource in many parts of the continent.

To address these gaps, I propose the following:

Investments in digital infrastructure, education and local start-ups to reduce dependency on foreign tech giants.

African countries must shape international AI governance forums. They must ensure policies reflect continental realities, not just western or Chinese ones. This will include using collective bargaining power through the African Union to bring Africa's development needs to the fore. It could also help with digital sovereignty issues and equitable access to AI technologies.

Finally, AI policies must embed African ethical principles. These should include communal rights and post-colonial sensitivities.