Facial recognition has a reputational problem. What was supposed to be an exciting new technology that could make our lives easier, safer and more secure has been shamed by reports that it has repeatedly failed when asked to correctly identify black faces. But Charlett N’Guessan, a pioneering software engineer from Ghana, has a different view of the ability of facial recognition technology, and her own products are designed to recognize African faces. Facial recognition tools created in the United States and other Western countries are often created by white people and trained on data sets of white faces, which means that when it comes to identifying people of color, especially blacks, there are more mismatches or outcomes.
When facial recognition goes wrong, the consequences can be severe, especially when the technology is being trial and used by police forces around the world. The Westernized tech industry often sees racial bias in facial recognition as a difficult problem to crack. But N’Guessan, 26, believes that as facial recognition technology matures, there’s no reason why it can’t recognize different faces, and that it’s just a matter of making sure the algorithm is robustly trained to use diverse data sets. The more models you train, the more faces your model will recognize.
Last week, Charlette N’Guessan’s work was officially recognized, making her the first woman to win the African Engineering Innovation Award. Founded and supported by the Royal Academy of Engineering, the award is the largest in Innovation in Africa. “Technologies like facial recognition must be based on African communities, and we believe their innovations will bring far-reaching benefits to the continent,” Rebecca Enon Chong, a Cameroonian entrepreneur and judge of the Africa Prize, said in a statement.
The motivation behind Charlette N’Guessan’s project is that research shows that Ghanaian financial institutions have a huge fraud problem, mainly due to cyber security. They spent a lot of money trying to solve this problem, so she and her team worked on a solution that could be integrated into their existing systems.