Some challenges in connection with the construction of a Massive Open Online Course on ethical and societal aspects of AI

The Swedish government is at present investing heavily in making Sweden competitive in AI (Artificial Intelligence). As a part of that effort, Lund University has been commissioned with providing a series of MOOCs (Massive Open Online Courses) on different topics regarding AI, including several topics dealing with Humans and AI. As a comprehensive university, Lund University has an advantage in being able to facilitate collaboration across academic fields to meet multidisciplinary knowledge needs in the wider society for the future of sustainable AI design, implementation, and use. The courses aim at a very wide audience but are particularly suited for those who did not major in AI related subjects at the university but who nonetheless have to deal with AI in their work. This resulted in a suite of courses of which one deals with societal challenges and ethical issues in AI. The course covers algorithmic bias, surveillance, democracy, responsibility, how to maintain control of intelligent machines, and conceptual issues regarding machine intelligence. In this chapter we present some challenges in connection with the making of this course, and how these challenges were overcome.