Algorithmic Injustice or Acceptable Imperilment. The Perceived Affordances of Documentary Film Coded Bias with regard to AI Awareness

No Thumbnail Available
Issue Date
Journal Title
Journal ISSN
Volume Title
Now that machines are able to process large amounts of data and are developed to intelligently produce results based on these data, issues have arisen concerning the inclusivity and fairness of the use of such machines. However, the general public is only sparsely aware of potential issues associated with artificially intelligent systems. In Coded Bias (2020), a Netflix-based documentary film, algorithmic bias is discussed as one of the harmful side effects of AI as an innovative digital tool. This research investigated the ways in which Coded Bias, as an informative medium, activates its audience's knowledge processing and, with that, induces increased awareness or attitude shifts with regard to algorithmic bias. Through a film analysis and interviews with viewers of Coded Bias as well as non-viewers, the affordances of the film were specified. Results show that the viewers report changes in the way they apprehend algorithmic bias but the nature of these changes depends on, among other things, their background and prior knowledge. Recollection, contextualisation, elaboration, and criticism are some of the notions that viewers relate to the changes they perceived in their view of AI and algorithmic bias after having watched Coded Bias. Consequently, it can be concluded that the film induces an increase in the viewers' awareness to some extent, at least temporarily. Yet the different ways in which viewers indicate their awareness and attitudes are affected suggest that the way in which each individual viewer perceives and interprets the information presented in the film plays a role in their processing of the information from the documentary.
Faculteit der Letteren