Can my algorithm be my opinion? : an AI + ethics curriculum for middle school students

Abstract

Children of today can be considered "AI natives." In the same way that children of the 90s were considered to be digital natives, children of the early 2000s and 2010s have grown up in a world where much of their access to information is mediated by artificial intelligence systems. Furthermore, we expect their futures to be increasingly affected by AI, as consumers and designers. For this reason, there is a movement to teach AI concepts to K-12 students. Drawing on a tradition of scholarship in Science and Technology Studies and a surge in recent research on the ethical issues associated with the construction of AI systems, it is clear that students not only need a technical education of AI, but an education that will allow them to become conscientious consumers and ethical designers of it. This thesis presents a set of standards which describe what every child should know about the ethics of artificial intelligence: that it is not an objective or morally neutral source of information and, given that, how to design AI systems with stakeholders in mind. It then describes a series of open-source, largely unplugged activities which address these standards by blending together ethical and technical content. Finally, it presents results from a pilot where students engaged with these activities. Findings about students' initial understanding of AI and the ethical dilemmas associated with it are presented, as are students' understanding after engaging with the curriculum. After participating, students moved from seeing AI as an objective tool to a tool that can be both objective and subjective. By the end of the curriculum, students were able to identify more stakeholders of technical systems and design their own systems according to the values of those stakeholders. This work shows that students can transform into conscientious consumers and ethical designers of AI.