Can chatbots and artificial intelligences revolutionise South Africa’s education system and help a generation of disadvantaged youngsters to achieve their personal dreams and ambitions and – in turn – grow the economy for others too?
It sounds a bit far fetched when you put it like that, but it’s pretty much exactly the system that’s currently being piloted in three township schools by a data organisation called Africa Success.
Africa Success has solid credentials. It’s been set-up by Yasaman Hadjibashi, an Iranian-born, German educated scientist who is currently the chief data officer for Barclays Africa who leads the banking group’s innovation agenda (and Barclays/Absa has been very innovative of late).
Speaking at TEDxJoburg this morning, Hadjibashi says that Africa Success came from a bit of research into the question of what makes young people from disadvantaged backgrounds successful later in life. Why is it, she asked, when young kids from the townships are incredibly ambitious at a young age do they fail to live up to their potential later in life?
By interviewing those who have “made it”, Hadjibashi says she identified five key things that differentiated them. Talent and skills and quality of education were obviously a major part, but so were participation in team sports, exposure to other cultures and independence or self belief.
But the most important differentiating factor was access to mentors who taught and inspired them.
It’s a data problem
Hadjibashi admits that to her, most problems are a data problem, but through Africa Success she’s applying her data science knowledge to the issue of providing enough mentors to teach and inspire each and every school kid in South Africa (and beyond).
Right now, the pilot schemes are based on matching human mentors with children in high school using WhatsApp. The plan, however, is to capture those conversations and use machine learning to begin to predict what sort of help children ask for from mentors. With enough data, it should be possible for the AI to know that its talking to someone who would benefit from an engineering mentor, for example, and being to anticipate questions, problems and motivational answers.
I asked Hadjibashi if she thought children would be able to tell that they weren’t talking to a human, falling in to the uncanny valley, so to speak. She said that she doesn’t think, ultimately, that it will matter. Getting the right information and support in the right language in a format that they’re comfortable with will outweight the disadvantages.
Africa Success is at a very early stage, but we’ll be following it closely to see how it develops.