Africa does not need to build the next ChatGPT to win the artificial intelligence race.
It needs to figure out where AI can make the biggest difference to the economies that already exist.
That was one of the more useful messages to emerge from an artificial intelligence conference in Johannesburg, where the conversation around AI extended beyond chatbots and humanoid robots to the less visible systems being deployed inside businesses and industries.
The distinction matters.
Much of the global AI race is focused on increasingly powerful foundation models, enormous computing infrastructure and consumer applications. Africa is unlikely to compete with the world’s largest technology companies on those terms.
But that does not mean the continent has to sit on the sidelines.
The AI that doesn’t look like AI
Some of the most valuable applications of AI may never become household names.
They could be systems that process invoices, predict equipment failures, optimise electricity consumption, improve supply chains, analyse financial transactions or help companies make better operational decisions.
South African technology company 4Sight Holdings, for example, is positioning AI around enterprise transformation, intelligent automation, data analysis and productivity rather than simply building another consumer chatbot. Its approach includes applications in smart mining, smart industry and smart manufacturing.
That model is particularly relevant to Africa.
The continent has enormous industries that remain relatively inefficient by global standards. Mining, agriculture, logistics, manufacturing, financial services, telecommunications and energy all contain processes where better data and automation could produce significant gains.
The opportunity is therefore not necessarily to invent a new category of AI.
It is to make existing economic activity work better.
Africa’s advantage could be application
Africa has a large and growing workforce, rapidly expanding digital markets and industries where relatively small productivity improvements can have significant economic consequences.
Consider a mine.
An AI system that predicts equipment failures before they happen could reduce downtime and maintenance costs. A logistics company could use AI to improve route planning and fleet utilisation. A bank could use machine learning to improve fraud detection and credit assessment.
None of these applications requires an African company to build a model larger than ChatGPT.
They require engineers, data scientists and businesses capable of adapting existing AI technologies to African operating environments.
That is a much more achievable proposition.
Local problems create local opportunities
Africa also has problems that cannot always be solved effectively by simply importing a technology developed for another market.
Language is one example.
African businesses and consumers operate across hundreds of languages, dialects and cultural contexts. AI systems that understand those environments can potentially improve access to education, healthcare, financial services and digital commerce.
The same applies to agriculture, informal commerce, public services and financial inclusion.
The companies that understand these problems may have an advantage over companies developing generic AI products for a global audience.
The opportunity is to build AI for African conditions, rather than necessarily building another general-purpose AI model.
Infrastructure still matters
There is, however, a danger in focusing only on applications.
AI needs computing power, reliable electricity, connectivity, data infrastructure and skilled people.
Africa’s AI opportunity therefore depends partly on the development of the infrastructure underneath the applications.
Data centres, cloud computing, fibre networks, local datasets and affordable access to computing capacity will become increasingly important as businesses move from experimenting with AI to deploying it at scale.
This is one reason the continent’s AI race should not be measured simply by the number of AI startups being created.
It should also be measured by how much AI capability is being embedded into the wider economy.
The real AI race
Africa may never need to win the race to build the world’s biggest AI model.
It needs to win a different race: using AI to become more productive.
That means helping a farmer make better decisions, a manufacturer reduce downtime, a bank detect fraud, a logistics company move goods more efficiently, a hospital manage information and a small business serve more customers.
The technology underneath those applications may have been developed in Silicon Valley, Shenzhen, London or elsewhere.
That is not necessarily a problem.
The economic value will increasingly come from how intelligently that technology is adapted, deployed and integrated into local economies.
Africa’s AI future may therefore look very different from the popular image of the AI revolution.
It may contain fewer humanoid robots walking across conference stages and more AI quietly working inside mines, factories, banks, farms, telecom networks and businesses.
And that may be a better outcome.
Africa doesn’t have to build the next ChatGPT to win the AI race. It has to become exceptionally good at using AI to build a more productive economy.
Sources: 4Sight Holdings; ITWeb; Johannesburg AI conference reporting.