Applying a principle of explicability to AI research in Africa: should we do it?