Artificial Intelligence (AI) is the current buzz in the tech world. We are going through a tech revolution that is changing the way we learn, organise, and process information.
There are even predictions of a time when AI-based machines will develop what experts call ‘Artificial General Intelligence’ and fears that in a (hypothetical) future called singularity, computer systems could develop autonomy and rebel against humans. In its extreme, AI-driven machines could enslave us or exile us to a remote planet!
AI is already solving some of the most complex human problems, including inventing groundbreaking vaccines like those used against Covid-19.
Unless the continent changes fundamentally, AI will not solve chronic issues in the continent, including corruption and mismanagement of resources, weak institutions, poor infrastructure, lack of transparency, inadequate legal and policy frameworks, inequality in access to resources, politicisation of everything, and dependency on foreign technology and expertise, among others.
In general, these problems are about governance. Technology and its derivatives only work if superimposed on stable governance structures. After all, technology is secondary to other basic human values.
First, AI will not solve Africa’s endemic corruption and mismanagement of resources. Even if we adopt such systems, our entrenched patronage structures could enable the state to manipulate AI tools to channel resources for political purposes. We are already good at data falsification.
A solution to this is to promote accountability systems with independent oversight for AI deployment in government projects.
Second, Africa’s weak institutions could hinder meaningful use of AI. In some countries, the Judiciary, for example, is highly politicised. The danger is that courts may side with the government in cases of AI misuse.
That many African countries have fragile election agencies is no secret. We are unable to manage even simple technology for elections. What about managing AI-driven voter systems transparently?
AI is still a human invention. Unreliable statistics agencies with poor or incomplete data could undermine AI outcomes. This is possible due to underfunded regulatory agencies, which could lead to an inability to monitor AI ethics and implementation.
Third, Africa’s poor infrastructure remains a problem. AI will not help much with the current unstable electricity grid in the continent, for example. AI servers and systems require much more stable power than is currently available in many countries. Africa already has slow and unreliable internet, which limits real-time AI applications.
In addition, the continent in general lacks data storage facilities and continues to depend on foreign servers, risking sovereignty. This also means weak cybersecurity frameworks, exposing the continent to hacking and manipulation.
Africa’s continued dependency on foreign technology and expertise could lead to the installation of ready-made AI models with embedded foreign biases. We already rely on foreign consultants, sidelining local talent. The result will be further loss of data sovereignty, with even sensitive national data stored abroad.With this, the risk of digital colonialism arises: global tech giants, like others in business, could become veritable agenda setters.
Fourth, there is a lack of political will for transparency. Without robust governance, what will prevent secretive decision-making where AI is used without public consultations or disclosures? What will prevent governments from blocking investigative journalism, making it hard to scrutinise government operations? We are already facing chronic opacity in public expenditure.
Fifth, Africa has inadequate legal and policy frameworks. This means outdated technology laws, which in any case are not designed for AI ethics and governance. There is weak enforcement of existing laws, with well-known political interference in regulatory bodies.
In Kenya, we are just beginning to put in place clear data protection policies. But this is not the case in many African countries, meaning citizens’ personal data are at risk.
This can be addressed by creating liability frameworks, spelling out the responsibilities for harm that may be caused by AI systems.
Sixth, there is inequality in access and benefits of AI. Even without AI, Africa faces the urban-rural digital divide, with ICT benefits concentrated in cities. AI will follow the same pattern. To make things worse, current AI systems are built without African language datasets, leading to linguistic and cultural barriers. There is also the exclusion of marginalised groups, where women, youth, and minorities are left out. The danger is that AI could be monopolised by elites, with AI tools in the hands of a small political-business class.
Seventh, AI may be used for propaganda, with automated bots pushing government spin. Across the continent over the years, some players have been adept at election manipulation. AI-driven micro-targeting of voters for disinformation has been made easier. There is no reason why corrupt officials will not feed biased or false data into AI systems.
If Africa does not address these governance issues, AI will not help much. AI will only improve societies that already have robust socio-economic frameworks. To paraphrase James Clear, the author of Atomic Habits, we do not rise to the level of our goals; we fall to the level of our systems.
