본문 바로가기
  • Home

University Students’ Perceptions of Gender Bias in AI-Based on In-depth Interviews

  • PHILOSOPHY·THOUGHT·CULTURE
  • 2024, (45), pp.319~361
  • DOI : 10.33639/ptc.2024..45.012
  • Publisher : Research Institute for East-West Thought
  • Research Area : Humanities > Other Humanities
  • Received : May 25, 2024
  • Accepted : June 24, 2024
  • Published : June 30, 2024

Kim Hyoeun 1

1국립한밭대학교

Accredited

ABSTRACT

The issue of AI bias is growing both domestically and internationally. The target groups affected by AI bias are typically gender and race. This study focused on the perception of gender bias, which is more influential in Korea, and conducted in-depth interviews with ten university students in their 20s who are about to enter the workforce, especially those with experience in A.I. development. Methodology: This study aimed to find out the perception of gender bias in A.I. and the current state of technology through in-depth interviews. To this end, we conducted a preliminary survey and a supplementary in-depth interview in stages. The results showed that university students were aware of gender bias in society and AI. However, they perceived that their personal experience of gender-related discrimination was independent of their development. Due to the nature of the field, recruitment is not affected by gender, but we did not exclude the possibility that an individual’s background and experience may affect the development process of AI, even if unconsciously. It was suggested that gender bias, if present, could be adjusted for by adding or weighting more data related to one gender. Most of the interviewees agreed that recent advances in bias mitigation techniques can be applied by applying existing techniques, suggesting that natural environments and training are needed to mitigate gender bias in AI. Based on these perceptions, this paper presents policy suggestions to materialize them.

Citation status

* References for papers published after 2023 are currently being built.