For four days in November 2023, Africa’s technology leaders convened in Addis Ababa, the Ethiopian capital, to discuss the continent’s digital future. AI was at the top of the agenda.
Though the leaders acknowledged that the rapidly growing technology “can stimulate economic growth” on the continent, they still made a case for regulation as part of a Continental Strategy.
While the conversation on AI regulation in Africa—like the rest of the world—is expected, a new study by Qubit Hub, an African-based AI research, innovation, and development lab, argues that focusing on building a strong foundation is crucial before introducing policies. The study claims policymakers should prioritize improving the state of the AI ecosystem. Countries like Mauritius and Egypt have dedicated national strategies to technology.
“Policy initiatives should be geared towards expanding computing facilities and internet connectivity, funding data centres, advancing the capabilities of Africa’s talent, and instituting policies that ameliorate data sets constraints,” the report argues.
Using a ‘four horsemen’ operational system framework, the report analyzed the crucial components of the AI ecosystem: data sets and data systems, digital infrastructure, talent, and markets.
At the foundation of any AI model is data. But Africa grapples with limited online data sets, according to the report. These limitations not only result in biased AI systems but also hinder the development of AI products for the African market since AI systems perform best when trained on data that is representative of the target user. While the report notes that there have been efforts to collect indigenous African data, it argues that there needs to be careful thinking about how this data is collected, handled, and stored to safeguard its authenticity.
AI models need data infrastructure to work, but Africa doesn’t have enough. According to the Data Center Map, Africa has 95 data centers out of 5,065 globally. Of the top 500 most powerful commercially available computer systems known to us, only one is located in Africa – in Morocco. The report makes a case for more investment in data infrastructure on the continent, noting existing efforts to bridge the gap: Africa Data Centre’s $500 million investment that will create ten hyperscale data centres sprout across 10 African countries within the next two years.
As AI adoption continues to grow on the continent, with AI-focused startups springing up, talent is needed to advance the design and development of solutions specific to Africa. The talent value chain in Africa is at the bottom of the heap, and though this has the potential for mass job creation, it poses unique challenges that may call for a rethink of African labour laws. One of the major issues to come out of Africa, and Kenya in particular, was the horrible working conditions that people hired to moderate the OpenAI platform and train its AI models, were subjected to.
The report also argues that for artificial intelligence to be properly maximized in the African market, there has to be more awareness of the benefits of the technology and its use cases should reflect the African realities. AI-focused solutions should address real-world challenges such as rural development, low literacy levels, and financial inclusion, among others. More importantly, efforts should be directed towards ensuring the commercial viability of these solutions. The big question is whether AI solutions can be profitable in Africa. The report proposes a two-way solution: innovation that reflects the socio-economic challenges of users and a focus on niche markets.