From the periphery, it’s easy to assume that Artificial Intelligence (AI) is still far removed from Africa’s realities.
But from simple fact-checking tools used by journalists in Mozambique and Uganda, to complex AI-driven revenue systems in South Africa and telemedicine in Rwanda, the technology has quietly become part of everyday life across the continent.
Yet, as experts at the Forum on Internet Freedom in Africa (FIFAfrica) 2025 in Windhoek, Namibia, cautioned, the lack of comprehensive AI strategies and policies across most African countries threatens to widen the gap between innovation and governance.
Many countries are relying on existing data privacy and computer misuse laws, which experts say cannot effectively govern AI. As of 2024, Africa was reportedly home to at least 2,400 AI startups, with a majority in tech hubs like Nigeria, Kenya, South Africa, Mauritius and Egypt, working on sectors such as agriculture, health, education, corporate services, manufacturing, e-commerce, business intelligence and gaming.
There are over 5,000 AI professionals on the continent, and African AI startups received at least $220 million venture capital funding in 2024. In Zimbabwe, AI-generated avatars like Alice are delivering digital news bulletins, offering a glimpse into the future of broadcast media.
But this seeming potential will remain just that without meaningful investments. During a panel discussion, Abudu-Sallam Waiswa, head of litigation, prosecution, and legal advisory at the Uganda Communications Commission (UCC), said that although AI has sparked excitement on the continent, its operations and regulation remain undefined due to the absence of tangible legal frameworks.
“Do we have sufficient knowledge, data, and statistics? Are we ready to legislate on this new animal called AI?” Waiswa asked.
He criticised the tendency by African governments to copy and paste AI policies from Europe, the US, or China, models that he said often fail to reflect local realities.
“For countries where less than 38% of the population has access to the internet, where most people survive on agriculture and don’t own smartphones, the priority should be building mindsets, research, and homegrown frameworks,” he added.
Waiswa argued that Africa’s approach to AI must start with deliberate investments in digital infrastructure, broadband accessibility, and equitable connectivity.
“Let’s prioritize building our people’s capacity, prepare their mindset to adopt AI carefully, and develop frameworks that speak to our context, not duplications of what worked in the EU or China,” he said.
He further urged the African Union (AU) to sponsor research and scholarships to harness AI’s potential while mitigating its risks. Without electricity and internet access for all, he warned, “AI may remain a far-fetched dream for many citizens in Windhoek, Harare, Kampala, or Accra.”
UNEVEN PROGRESS ACROSS THE CONTINENT
A new report by the Collaboration on International ICT Policy for East and Southern Africa (CIPESA), ‘State of Internet Freedom in Africa 2025: Navigating the Implications of AI on Digital Democracy,’ echoes these concerns.
It found that countries with higher ICT development indices, such as South Africa (85%), Tunisia (79.6%), and Egypt (77.9%), also lead in AI preparedness. Meanwhile, Mozambique (32.4%), Uganda (42.4%), and Cameroon (46.3%) rank among the lowest.
Despite these disparities, AI has shown potential as an equalizer. In Kenya, the Umati Project utilises natural language processing to monitor hate speech and disinformation online, while Ghana’s Fact-Checking Coalition, Ethiopia’s HaqCheck, and Nigeria’s iVerify employ AI tools to combat misinformation.
Zimbabwe recently experimented with its first paperless, AI-powered trial. Governments across the continent are also exploring AI in surveillance, law enforcement, and service delivery.
Ghana has introduced biometric-based border management systems with e-gates, Kenya plans to upgrade its airports with AI-powered security and biometric gates, while Rwanda aims to develop 50 AI applications across key sectors within four years.
According to the Oxford Insights Government AI Readiness Index (2024), Egypt, Mauritius, South Africa, and Rwanda are Africa’s front-runners, while countries such as Chad, Eritrea, Burundi, and South Sudan remain at the bottom.
PROMISE AND PERIL
AI’s promise lies in its potential to transform Africa’s digital, social, and economic landscape, from e-learning and data analytics to diagnostics, agriculture, and governance.
The rise of Generative AI has further accelerated this momentum, with its potential for innovation in content creation and data-driven decision-making. But as Wairagala Wakabi, executive director of CIPESA, warned, enthusiasm alone is not enough.
“Governments are hardly doing anything in terms of investments, either in innovating, or in scaling AI, innovations or documenting usage or taking them up significantly. In the area of policy development, just a handful of countries – less than one third – have come up with policies. No country has an AI Act. Even those with policies, hardly has there been multi-stakeholder consultative policies. A lot of time is top-down,” he said.
Wakabi noted that for Africa to truly benefit from AI, investment in research, policy development, and stakeholder engagement must be prioritized. He said policy development should not be the main thing, but rather investment in knowledge and awareness capacity to develop proper oversight for greater productivity.
For Dr Eileen Carter, national coordinator for Human Rights and Developing Technologies, AI policy must be grounded in human rights principles.
“There’s no need to rush or superimpose policies from elsewhere,” she said. She said that approach risks creating tools of oppression rather than empowerment. AI must serve people, not violate their rights.
DIGITAL COLONIALISM
A new concern, the report noted, is that the African continent is being used once again for AI experiments, especially by China. Kenya, Rwanda and Uganda have been cited as countries where AI experiments are being tested on their own citizens.
In 2021, the report notes, Zimbabwe signed a deal with Chinese firms – CloudWalk Technologies and Hikvision, to hand over biometric data to train facial recognition systems, which has been linked to a “chilling effect” on dissent, with citizens self-censoring due to fears of being monitored.
Also, recent research highlights that state surveillance in Zimbabwe and Uganda has had a significant chilling effect, resulting in self-censorship, directly restricting their own expression; an unwillingness to engage with individuals or organisations believed to be subject to surveillance; and the erosion of trust, affecting individuals’ ability to form and maintain relationships, build networks and organise politically.
About the Author
