ABSTRACT
The chapter focuses on how a techno-optimistic stance that technology is improving society is upheld, in spite of evidence that AI systems can lead to algorithmic bias. The chapter discusses how people working on AI in hiring construct algorithmic bias as solvable. It is suggested that algorithmic bias is due to data, which means that the AI learns bias from humans. In order to fix this bias, it was suggested to fix the data, to fix the human raters who produce data and to fix the algorithm. Some interviewees expressed what is called techno-hesitation. This stance entailed that decision-making power rests with humans and that due to reputational risk from negative media reporting and lawsuits, using AI in hiring is avoided. However, this was a temporary stance in that most people expected that after a tipping point is reached and societal acceptance increased, AI will be used in hiring regularly and routinely. The chapter shows how algorithmic bias is constructed as something that can be fixed, which functioned to support a techno-optimistic worldview.
