Why Africa isn’t ready for artificial intelligence and its disruptions

AI is expected to lead to a reduction in the size of the workforce as it automates a lot of human work. FILE PHOTO | SHUTTERSTOCK  

By MICHAEL MICHIE More by this Author

Is Africa ready for artificial intelligence? The short answer would be No. With a lot of technological advancements, Africa has been fortunate not to have gone through the...

evolutionary cycle that comes at the point of a major shift in technology.

Some parts of Africa will not know the pain of moving from 2G networks to 5G, the pain of cost and infrastructure.

Africa can, as it has with most technologies, leap to adopting the newest or relatively newest technology and scale itself a lot faster.

What makes artificial intelligence significantly different? Artificial intelligence (AI) was the buzzword of late 2022 in the tech world and it continues into 2023.

Unlike all other technologies, this is one we should not import. There are too many black boxes around AI that would require significant regulation.

These black boxes are a result of the nature of how current AI systems are being trained, the data training them and their use.

When it comes to data that train AI it is common to want and attempt to remove biased data.

However, humans have natural biases and some of them we are not aware of. These biases vary across cultures and as such it would be more responsible to train AI and have it developed by the people who would consume it.

Without such an approach the black box becomes bigger and inaccuracy increases significantly. We should not attempt to use AI to eliminate biases at the first attempt but to fully expose any existing biases that are unknown.

AI is expected to lead to a reduction in the size of the workforce as it automates a lot of human work.

With the growing population in Africa and large numbers of unemployed, a lot of effort is required to formulate how AI will fit in the workforce without bringing a significant increase in unemployment.

Programmes to reskill those who might be affected to be able to work with AI in a paired work structure should be developed.

Skilling of the unskilled to utilise AI or to move to industries that are not likely to be impacted heavily by AI is needed as well.

Infrastructure for AI is important to ensure safety for its direct and indirect users in Africa. The infrastructure includes the hardware required to train these models as well as the data.

A lot of the data in Africa is yet to be digitised in a form that can be used to effectively train AI.

The underlying process to resolve this also requires infrastructural changes in data capture in the private and public sectors to ensure data collected is not lost and is stored electronically.

Other challenges of importing AI would all be resolved by establishing strong frameworks for its use, development, and training, creating regulations, frameworks, policies, and strategies around how data for training would be collected, anonymised, and used to train models.

AI is good for Africa, but AI should be built by Africa for Africa.

Michael Michie is the co-founder of Everse Technology Africa.