Bias in AI and Machine Learning: The Impact of COVID-19 in African Healthcare Communities

AI technology has become increasingly involved in a plethora of societal functions in recent years, but racial bias in AI algorithms has revealed a dangerous trend. With the rapid advancements in technology in general as well as artificial intelligence algorithms, bias is unknowingly developing in these algorithms due to the lack of attention towards it. However, recent efforts have been made to first recognize that bias exists in these algorithms as well as strategies to eradicate it. The implications of the research performed in this analysis go much further than a simple moral obligation to promote inclusiveness for marginalized groups in society; racial bias in AI algorithms has the potential to involve life and death consequences. Specifically in the provision of health, an unbiased algorithm may inherently contain bias due to factors outside of the algorithm itself. It is important to use diverse data sets in our algorithms to ensure that the data does not contain bias. Using a data set which is not diverse may lead to the algorithm developing bias over time, which may cause adverse impacts on patients. In addition, we will discuss how bias affects Africa in comparison to more developed countries. We will look into the future of how we can eliminate bias in artificial intelligence and advance the provision of health more equitably across the global community. Based on findings showcasing examples of racially biased AI technology used to combat the COVID-19 pandemic, efforts currently being taken to eradicate racial bias in AI are highlighted along with a discussion of future actions that should be performed. With more federal regulations surrounding AI algorithms along with an emphasis on promoting diversity in the personnel and data of the AI community, particularly in Africa, the future of AI can be one free of racially biased tendencies.