Compromised Agency and Ethical AI Development in Africa

Today I woke up thinking about the increased global push back on AI and large tech companies. While many nations consider AI’s enormous abilities and possibilities, there is an increased awareness on how this may mess up with agencies, especially “regarding cases of alternative ways of knowing and doing science such as civic, and feminist sciences, among other” as highlighted by Max Liboiron (2017) in his article, Compromised Agencies: The Case of baby legs.

In Africa, the concept of agency is very much present, and is evident with increased agitations from some groups like tech workers. They have so far lodged several lawsuits seeking for fair work conditions from large tech companies and their subsidiaries who are outsourcing labour from African and other global south countries. However, there is a disturbing silence among the technology developers and workers in high end positions in large tech. One may ask, is it that the tech being developed in Africa is 100% ethical? Or do we have workers whose agencies have been critically compromised?

I am a picker of the most unimportant statements in informal conversations. In the past I have heard statements like “Company X takes care of you, as long as you work for them”, … “Sometimes you’ve got to think of where you get your food, and overlook things” … “You have to look at the bigger picture” … “We have made sure that most of our tech developers are Africans and we have social scientists focusing on the ethical issues”. These statements are so innocent, and some of them reflect inclusivity to an innocent on-looker, yet they are packed with compromise – but only from the weak. What I am saying is that focus on innocent or pure agency is an exercise in futility, especially because it “often glosses over the constraints placed on agents, particularly within asymmetrical power relations”. Most tech developers in Africa working for large tech companies have in many occasions admitted that they serve the objectives of their masters.

Let us consider a cases young person, who has been unemployed or underemployed for years. Then they hit a well-compensated position in large tech, where some level of agency is placed on them. In addition they are exposed to the ambit of large tech, and access of other benefits that places them high above their peers. I listen to young people a lot, and one once said to me “I would kill to defend a company that pays me well”. Such a person has limited ability to bring to light ant cases of AI malpractice, or any ethical issues that put the populations they are serving at risk. Even if they make an effort to exercise some agency, this agency is never pure, it is compromised as it is “exercised within asymmetrical power relations”.

Sometimes I wonder, how it would it be like if we were all able to compromise something? Whether tech company owners, practitioners or low-end workers? That way, compromise in tech is not left as the place for the weak, and those who are not in power nor economically endowed, rather as a place for intentional thinking and considering how diverse scientific and non-scientific knowledges actors might compromise to ensure that alternative ways of knowing and thinking are allowed to emerge in large tech, including AI. This way, our way of compromising leads to development of AI that serves all classes of people in Africa.

Scroll to Top