Datafication in Human-Machine Communication (HMC) most often relates to data used for training, processing and presenting the AI of the system on the interface and is associated with problems of privacy, discrimination, and representational biases. This chapter introduces to these datafication issues and go into depth with the challenges of datafication and representation biases from a different perspective: when the nonbinary representation ideal meets the preferences of users of different genders. Through gender theory and a quasi-experiment case study of preferences of 12 binary and nonbinary users in relation to voice-controlled assistants as a growing user interface within HMC, the chapter exemplify questions and issues within datafication and how to study them. All participants in the tested groups exhibit gendered descriptions of the voice assistants, even when trying not to. Seven out of twelve preferred the voice assistants better when the perceived gender of the assistant was congruent with their own gender identity. The study shows how having an ideal of nonbinary representation faces challenges when it comes to both the datafication of such non-binary representation in voices (e.g. uncanny valley and contextual awareness issues in HMC) and the preferences of different genders due to the historical dominating binary interpretation culture.
|Title of host publication
|Handbook of Human-Machine Communication
|Place of publication
|Published - 2023