In 2021, led by Sandra Nabulega, Pollicy began to research the impact of Artificial Intelligence on African women. Moved by the absence of data and information on this topic, we decided that this was a project worth embarking on. In places where AI had already been deployed and in use, there were numerous examples of how AI inadvertently discriminated against women. Infamous examples were Amazon’s recruiting tool that ranked the CVs of women lower than men’s, Microsoft’s short lived chatbot Tay, and the proliferation of female-gendered chatbots and virtual assistants, which sparked concerns about the societal impacts of casting women into subservient roles (outlined in UNESCO’s I’d Blush If I Could report). Worse, Joy Buolamwini and a team of other researchers revealed that AI facial recognition software is terribly inaccurate at identifying the faces of Black women. Results show a 50–50 chance of accuracy.
We wondered if there were similar experiences for African women and if AI discrimination was a thing we needed to worry about on a continent where gender parity stands at 0.58 out of 1. We were also curious about the experiences of women working in AI. We soon realized that research of this size could not be adequately contained or addressed in just a single report and thus this research is split into three separate chapters.
The first report, Engendering AI explored more general themes related to AI in Africa: the hindrances, the supporting institutions and various definitions of AI. It also provided a glimpse into what this project was designed to do: explore the impact of the interactions between AI and the unique and intersecting identities of African women as African, as women, and as African women. It briefly touched upon colonization and the environment in which AI itself was conceptualized, and how the influences of primarily White, male founders define what intelligence means.
The second report, Encoded Biases & Future Imaginaries, sought to go deeper into the impact of AI on African women. After all, AI is all around us today. It is in our phones, on our social media apps and on our lending apps. We are writing and designing with it, scuttling elections with it. Our governments are launching surveillance systems with it and if we try to migrate to Western nations, our visa applications are being analyzed by these algorithms. Our major challenge here was determining how to present the impact of AI both positively and negatively. Even some of the most benign uses of AI could very easily be used for surveillance, to breach privacy, cause indignity and further marginalize already marginalized groups. Some questions were difficult to answer: Was the use of AI to track victims of human trafficking worth the risk of creating data collection systems which would or could fall into the hands of anti-democratic institutions? Was the use of AI in classrooms worth the risk of unemployment to teachers who are poorly compensated? These remain questions that are difficult to address.
For our final report, we reached out to African women: those who are AI developers, people who fund AI research, teach AI, work on AI ethics, or use AI in their work. A 2021 World Economic Forum report found that women comprise only twenty-five percent of data and AI positions in the global workforce. This statistic became even more apparent when we began to conduct our interviews last year; we struggled to find respondents, especially in fields such as nuclear energy and climate change. On the other hand, we more readily found respondents working in the crucial fields of health and agriculture. Our respondents revealed to us their aspirations, thoughts, and desires. They desire to see more women work in AI, and many of them support women in their various capacities to join.
The goal of this third chapter was to explore their lived experiences. African women in AI have pivoted many times, persevered and have received many pleasures. This paper is surely a fitting end to this “book,” but it is only the beginning of the immense effort needed to ensure that the needs and voices of women in Africa are centered in the creation and dissemination of this technology.
Written by Favour Borokini (Data and Digital Rights Researcher at Pollicy)
