In Africa, the technology of facial recognition is not commonly used, partially because the technology available until now has not been able to recognise and differentiate between black people’s faces. The best Western facial recognition technologies have been checked by the US government and found to misidentify Black people at rates five-to-10-fold superior to white people. There was a strong problem with the racial disparities in biometric artificial intelligence technology which is a foundation for these systems: the use of data settings made up entirely of white faces trains them.
In 2018, four software engineers in Ghana set up a company to deal with the widely available software of facial recognition. The research they carried out revealed that the banks from Ghanaian are affected by wider identity fraud and cybercrime and spend close to $400 million annually on customer identification. The group created its own facial recognition programme, BACE API, using artificial intelligence and was led by Charlette N’Guessan, an engineer originally from Côte d’Ivoire. Led. Contrary to western developers, they training BACE API with a broader range of data sets and a broad representation of the local Black African faces. One of the main strengths of BACE API is the remote control of a user’s identity, particularly useful to financial institutions and other client-oriented industries whose operations involve ID checks. It can be implemented in current applications and systems and needs no special hardware. The use of live images or short videos to decide if images are real people rather than existing pictures is a further distinguishing feature of the BACE API.
For the 2020 Africa Innovation Innovation Engineering Grant, which the UK’s Royal Academy of Engineering awards annually, the community software overcame other competitors. N’Guessan, 26, was the first woman to win the competition, winning £25,000 in prize money.
Inputs from QZ
