Blueprint: Artificial Intelligence for Africa

"Currently, no African country is ranked among the top 10 countries expected to benefit most from A.I. and automation. In contrast to the leading A.I. nations such as China, the U.S. and others, African countries have been in general rather slow in the adoption of A.I. technologies. However, at Smart Africa, we believe that Africa cannot and should not be left behind!"

Executive summary

Artificial Intelligence (AI) is a technology that can propel developing economies on a trajectory of sustainable development. This is however achievable only through the application of a coordinated, integrated approach to the development and implementation of AI strategies by Member States. 

The African continent’s intent is on leveraging the digital transformation process to achieve sustainable growth and development, as evidenced by the SMART AFRICA initiative. One of the main elements of this digitalization thrust relates to the application of AI technology. However, to date, a clear strategic and practical framework for the use and governance of such technologies, has not yet been articulated or documented. This AI Blueprint therefore lays the foundation for the development of an AI strategy and presents key elements and considerations to be borne in mind for the formulation of such. 

The overall objectives of the blueprint are: 

• to outline the most relevant opportunities and challenges of the development and use of AI for Africa and how to address them. 
• to make concrete policy recommendations to harness the potential and mitigate the risk of AI in African countries.

Against this background, Smart Africa created an AI working group with experts from Member States, the private sector, international organisations, academia and entrepreneurs, for guiding the development of an AI blueprint for Africa. The Republic of South Africa, one of the Smart Africa Member States, has committed to champion the development of this blueprint as part of their flagship project “4th Industrial Revolution: innovation and artificial intelligence” within the Alliance.

AI refers to general-purpose technology

As presented in Chapter 1, AI refers to general-purpose technology. Like other gene[1]ral-purpose technologies did in the past (e.g., the steam engine or electricity), AI is trans[1]forming our world, our society, and our industry. Such advancements have become possible due to the concept of machine learning. As a result of this development, there is great versatility in the application of AI and as such, fields of medicine, agriculture and meteorology have all benefited.

More developed economies like the United States of America, the European Union and China are at the forefront of the development and implementation of AI policies and strategies. Africa has recently commenced its journey in respect of leveraging AI and is well positioned to develop a coordinated strategic approach to the use of AI at the national and regional levels.

Lessons learnt on AI governance and strategies

As intimated in Chapter 2, Africa can leverage lessons learnt from the experiences of other countries which have undergone the development and implementation of AI strategies. This AI Blueprint therefore proposes the application of 6 key principles to the development of an African AI Strategy, notably: 

• The inclusion of AI as part of a wider national strategy
• Balancing the development of an AI enabling environment against ethical, legal and governance considerations 
• Underscoring the importance of both the process and the plan
• Focusing on action 
• The application of an inclusive approach to AI- by the people, for the people 
• Leveraging the national AI strategy as a tool for communication 

Chapter 2 further explains and concludes that flexibility should represent a core pillar in the African AI strategy, particularly in respect of governance, given the innate fluidity of AI technology. The strategic governance framework should be sufficiently malleable to allow for the implementation of both proactive and timely responsive measures to cater for developments in AI technologies. 

Based on the premise that an all-inclusive approach should be taken to the development of an AI strategy, Chapter 3 explores the framework for national AI strategies in Africa. It discusses and highlights the relevance of the 5 pillars of an AI national framework.

AI strategic framework for Africa

These pillars are noted to be: 

• Human capital, which underscores the importance of educational development in respect of AI. This pillar addresses the relevance of enhancing the proficiencies, competencies and understanding of individuals in respect of using and developing artificial intelligence solutions.
• From Lab to Market initiatives that foster research, development, innovation, and commercialisation. 
• Networking, cooperation and collaboration, in pursuit of joint partnerships across private and/or public sectors in an effort to favourably impact the uptake of AI. 
• Fostering the development of digital and telecommunication infrastructure to pursue efficient data collection and usage. 
• Effective regulation that is premised on the infusion of ethics and international best practices.
• Governance and Ethics for AI

Chapter 4 presents an analysis of governance and ethics within the context of regulating AI. In emphasizing the importance of digital governance, this chapter compares various regulatory approaches. The analysis identifies that an adequate legal framework must consider AI applications, data, ethics, the establishment and operationalisation of an enabling business environment, the multiplicity of regimes and industries which must be catered to and the application of soft law or hard regulation, as applicable. The use of hard regulations is advocated in situations where soft law or the marketplace itself, cannot resolve market concerns. The application of hard law is most applicable in instances that interplay with:

• The regulation of copyright/patents
• The promotion of investment, protection of intellectual property and regulation of accountability systems 
• Unfair competition laws or other similar regimes

Soft regulation is noted to establish expectations that are not enforceable by government and includes instruments such as professional guidelines, codes of conduct/codes of ethics and international best practices. Chapter 5 discusses the practical use and application of AI to real-life, noting key sectors which can potentially benefit from the leveraging of this general-purpose technology.

AI to real-life – concrete use cases

Key sectors and areas of focus discussed are agriculture, education, health, financial services, energy, transportation, and climate change. This Blueprint proffers that a clear roadmap is required to ensure that real benefits are derived from the application of AI technology. Accordingly, Chapter 6 sets forth a roadmap for Africa in this regard. The roadmap proposes to include: 

• Collaboration in the formulation of policies, ensuring that there is harmonization, inclusivity and international best practices infused therein. To achieve this, it is proposed that a coordination mechanism for AI policy at regional level is established and similarly, peer to peer exchanges and benchmarking mechanisms are to be instituted. 
• Creation of an open data environment for Africa, leveraging the lessons learnt from other initiatives (e.g. the European Commission). This element advocates for the establishment of AI-ready data as a public asset and the normalizing of the concept of open public sector data to reduce entry barriers and promote AI innovation. 

• Common Computing Infrastructure to host, process and use data, thus cultivating a culture to foster the development of data lakes, enabling data analytics and machine learning.

   

Data is a critical AI asset

Data is a critical asset for developing AI. In this blueprint we cover several aspects in detail, from how to accelerate the availability of data assets by leveraging the usage of non-traditional data or facilitating sharing data from public to private, and how to define data strategies to cover the regulatory aspects in data privacy and accessibility. In the end we extend our considerations to a wider regional approach in sharing data and infrastructures. 

Chapter 7 summarises a list of recommendations, which take into account the potential of lessons learnt from several countries for the development and use of AI for sustainable development, sector priorities for AI in the African continent and the landscape of use cases analysed in chapter 5. 

In conclusion, AI presents unique opportunities for the African continent. In countries with high uptake, AI has proven to improve service delivery, efficiency and effectiveness both in the public and private sectors.