Africa’s AI future will be strongest if it is built on the continent’s own data and priorities and taught in its own languages, participants heard during Huawei’s seventh yearly ICT Editors Xchange held in Johannesburg last month.

Intelligence built on Africa’s terms is not only possible but already under way.

The event featured speakers from Huawei, academia and business who addressed African AI, local data, indigenous-language education and the cloud infrastructure needed to support it under the theme ‘Africa’s Intelligence, Africa’s Terms’.

Huawei South Africa deputy CEO Charles Cheng said that, as the next chapter of Africa’s digital transformation would be driven by cloud, AI, connectivity, secure networks, data infrastructure and talent, the continent should define the terms of its own intelligence, rather than inherit them.

“African intelligence [on] African terms means systems that are trained on our realities and designed for our needs. It means innovation that works with small businesses, schools, hospitals, researchers, entrepreneurs and public institutions.

“It means ownership of the platforms and autonomy over the questions we ask, the problems we choose to solve, the data we use, the languages we include and the value we create,” he said.

AuraaAfrica founder and CEO Gift Lubele agreed that AI must be trained on African data to serve African users better, highlighting that 94% of all AI music models are trained on Western data.

“African data is 0.04%, not even 1%. If we do not include ourselves in these AI models, we cease to exist,” Lubele said, urging a move by the African continent away from consumption to rather becoming a creator of data.

He founded AuraaAfrica to help close that gap and recently won the Grand Innovation Prize at Huawei’s Code4Mzansi developer competition, which attracted 1 040 entries.

Lubele warned that the opportunity is time-bound.

“We have a window to catch up with the world, and that window is closing if we do not use it,” he said, noting that merely consuming AI tools sends money and intellectual property out of the continent.

Digital futurist Nicky Verd argued that Africa’s AI advantage lies in its ability to adapt quickly rather than being constrained by legacy systems, warning that the greatest risk is complacency, not technology.

The University of Pretoria’s Dr Olaperi Okuboyejo pointed out that mainstream AI tools still perform poorly in African languages, adding that AI can support learning when it is delivered in the languages African students actually speak.

Unpacking a paradox in South African education – a record 88% matric pass rate and the largest- ever cohort, alongside a deep early-learning crisis – she said that 81% of Grade 4 learners in South Africa cannot read for meaning in any language, which is partly linked to a language switch that pushes children into English instruction when only a small fraction speak it at home.

Okuboyejo highlighted her lab’s work on indigenous-language datasets and pan-African language models.

Meanwhile, Huawei Cloud South Africa cloud solutions sales director Calvin Huang explained the cloud and computing layers that African AI ambitions depend on, tracing a rapid shift from generative to agentic to physical AI, citing Huawei founder Ren Zhengfei’s view that AI may be “the last technological revolution in human society.”

He said the economic opportunity is equally significant, with AI investment projected to reach $810-billion by 2028.

However, African users remain “more like consumers of AI, not builders.”

“South Africa still faces challenges around deployment, cost and data sovereignty. Huawei is addressing these challenges through open AI systems, local infrastructure and industry partnerships.”

Huawei’s AI infrastructure is now live in its Johannesburg region, the same infrastructure Lubele is using to train his music models.

Huang pledged that Huawei would be “a closer, faster and more affordable infrastructure provider to South African customers” and “a long-term partner in this journey.”