We are told that Artificial intelligence (AI) has the potential to be a powerful tool for advancing democratic concerns and human rights across Africa. Yet, there are also early indicators that AI could undermine democratic institutions and processes, especially if these technologies prioritise colonial-capitalist development trajectories. Scott Timcke looks at some of the issues at stake.
By Scott Timcke
Fortunately, there are institutions, universities and think tanks across the continent where people are thinking about Artificial Intelligence (AI) at the nexus of democracy, technology, and economic issues. This post briefly outlines some of the key challenges that are relatively under-discussed when covering efforts to preserve and extend democracy in Africa.
Simplified, democracy is a system of government in which power is vested in the people, who rule either directly or through elected representatives. Other core principles include freedom of expression, consciousness, and assembly, as well as free and fair elections with universal suffrage.
The current orthodox understanding of modern democracy was conceptualised and designed during industrialization, with some extension and modifications through public administration and the welfare state after the Great Depression. The wave of decolonization also included expressions for liberation, enfranchisement, representation, and independent self-rule to overthrow colonial occupation. African people, including its intellectuals, were active participants in these debates and developments. Some embraced the democratic idea, others emphasised aspects like economic self-determination, while others thought liberation could be found through other methods and frameworks.
Nevertheless, the variety of responses, nearly three generations of Africans have fought hard for democratic representation; they have done the hard labour insisting on equality before the law; and they have struggled tirelessly to create and defend democratic elections. Given this history, it is imperative that African people be cautious to ensure that AI does not undermine this political investment.
Mineral, metal and data extractivism
Africa possesses substantial reserves of minerals and metals that are essential for manufacturing AI systems, including cobalt, tantalum, and gold. However, mining has often exploited African resources without providing sustainable benefits for local populations. It is also true that mining can bring ecological destruction, destroying the quality of life for directly impacted communities. Yet, African governments have significant leverage in the responsible sourcing of minerals which are key for ethical AI, both in Africa and beyond.
On the topic of avoiding rampant extractivism, there are valid concerns about a ‘digital scramble for Africa’ where foreign headquartered multinational technology firms extract and monopolise data from the continent, concentrating AI capabilities in the Global North or China. This data extractivism mirrors historical patterns of resource extraction from Africa by colonial powers. Domestic African data governance and AI development capabilities are critical to ensuring the continent shapes its own digital future.
Although not the only body, the AU is well positioned to remind governments that the development of AI systems ought to prioritise democratic governments over brute capitalist extractivism, whether it be for minerals and energy or whether it be for data. Doing so can help mitigate or avoid ‘digital colonialism‘ where multinational technology firms extract African data without permission or benefit for local populations.
Labour protections for click-workers and remote workers
Turning to labour concerns, across the continent there are indicators that the future of AI related work in Africa will likely be click-workers doing data cleaning and labelling. As infrastructure expands, and confronted with poor employment prospects, young people across Africa, will likely participate in the global gig economy, like completing microtasks on online platforms. This kind of granular sub-contracting work presents significant challenges, for example, low wages, repetitive tasks, the real prospect of AI managers, and the lack of access to essential social safety nets. These undignified working conditions ought to prompt African governments to consider adjusting labour protection policies to specifically respond to this pattern of precarious work while also making unionisation efforts easier.
Beyond the realm of AI, radical African voices will likely need to experiment with designing and implementing industrial policies that are labour-absorbing. These can contribute to social well-being. By proactively shaping the future of work, African governments should be forced to bridge the protection gap for vulnerable workers and build a more resilient and equal economy.
Revisiting neoliberal assumptions
Mainstream economics has become closely associated with neoliberal policies favouring free markets and limited government intervention. However, the neoliberal approach has failed in addressing issues like inequality, climate change, and economic resilience.
As Harvard economist Dani Rodrik recently wrote in an essay published by the neoliberal IMF, “The neoliberal policy paradigm favors expanding the scope of markets (including global markets) and restricting the role of government action.” Importantly, he adds that “Today it is widely recognized that this approach failed in a number of important respects. It widened inequality within nations, did little to promote the climate transition, and created blind spots ranging from global public health to supply-chain resilience.”
There now exists a window to develop and deploy new approaches to tackling the major challenges Rodrik lists. This includes the assumptions around AI governance and the role of the market. The shifts in capitalist-globalisation today might provide space for new models allowing more democracy in national policies that are context-specific and take into account a country’s institutions, industry, demographics and geography.
As an important body, the AU could play a pivotal role in promoting a democratic use of AI across the continent. Given the global nature of AI systems, we must urge multilateral cooperation in developing shared ethical AI frameworks. Already existing AU bodies have established critical working groups to lay out guidelines for accountable AI aligned with democratic principles.
Even so, there is scope for labour-led initiatives to develop imaginative policy solutions tailored to Africa’s socio-economic context that leverage the cost of AI while regulating risks. Ideally African countries should act together to retain the policy space needed to protect the rights of workers to shape their collective digital futures.
Scott Timcke is a Senior Research Associate at Research ICT Africa and a Research Associate with the University of Johannesburg’s Centre for Social Change. His second book is Algorithms and The End of Politics (Bristol University Press).
Featured Photograph: Digital training in Ghana (Christian Yakubum, 17 April 2024).
I think a lot about this article is quite a moot point. beyond my skepticism over the capacities and scope of the AU. AI requires such a tremendous amount of energy, and most of the physical infrastructure is already in the US, and then theres a bit in the EU nations and England. and in the us, ai and data centers are quite a heavy strain on the energy grid.
Further, i wonder how you square the circle of technocracy, welfare systems and democracy. technocracy, especially by unelected officials, is antagonistic to democratic imput. In fact, much of what has followed WW2 has been an expanded role of administrative management by the state and regulatory bodies. Further, a lot of neoliberalism actually requires and calls for a very active state, to simulate, create and facilitate markets, either by enforcing markets, or public-private/private-public partnerships. Inequality is already baked in and expected, and in terms of the problems the neoliberals identified, it has been quite successful at its goals — creating new markets, boosting profitability, maintaining and further entrenching marketed relationships etc.
Lastly, much of the emphasis on these digital technologies in Africa has been on the standpoint of software, not much on hardware, nor is there sufficient energy infrastructure to actualize these developments. Africans can be used as cheap coding labor, much like in india, in order to undercut the market in the global north, and thats also a bit of the hope with ai seemingly, to undercut the market of software devs in the global north/to further weaken their bargaining power.