To read this content please select one of the options below:

Gender Dynamics in Human-AI Role-Taking

Jenny L. Davis ^a

Daniel B. Shank ^b

Tony P. Love ^c

Courtney Stefanik ^b

Abigail Wilson ^b

^aThe Australian National University, Australia

^bMissouri University of Science and Technology, USA

^cUniversity of Kentucky, USA

Advances in Group Processes

ISBN: 978-1-80455-154-7, eISBN: 978-1-80455-153-0

Publication date: 27 October 2022

Abstract

Purpose

Role-taking is a basic social process underpinning much of the structural social psychology paradigm – a paradigm built on empirical studies of human interaction. Yet today, our social worlds are occupied by bots, voice assistants, decision aids, and other machinic entities collectively referred to as artificial intelligence (AI). The integration of AI into daily life presents both challenges and opportunities for social psychologists. Through a vignette study, the authors investigate role-taking and gender in human-AI relations.

Methodology

Participants read a first-person narrative attributed to either a human or AI, with varied gender presentation based on a feminine or masculine first name. Participants then infer the narrator's thoughts and feelings and report on their own emotions, producing indicators of cognitive and affective role-taking. The authors supplement results with qualitative analysis from two open-ended survey questions.

Findings

Participants score higher on role-taking measures when the narrator is human versus AI. However, gender dynamics differ between human and AI conditions. When the text is attributed to a human, masculinized narrators elicit stronger role-taking responses than their feminized counterparts, and women participants score higher on role-taking measures than men. This aligns with prior research on gender, status, and role-taking variation. When the text is attributed to an AI, results deviate from established findings and in some cases, reverse.

Research Implications

This first study of human-AI role-taking tests the scope of key theoretical tenets and sets a foundation for addressing group processes in a newly emergent form.

Keywords

Acknowledgements

Acknowledgments

Research was sponsored by the Army Research Office and was accomplished under Grant Number W911NF-19-1-0246 to Dr. Daniel B. Shank. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the Army Research Office or the US Government. The US Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation herein.

The authors would also like to thank members of the Humanising Machine Intelligence Project at the Australian National University, especially Dr. Bianca Baggiarini, Dr. Claire Benn, Professor Toni Erskine, and Dr. Jennyfer Taylor who provided feedback on earlier drafts of this paper.

Citation

Davis, J.L., Shank, D.B., Love, T.P., Stefanik, C. and Wilson, A. (2022), "Gender Dynamics in Human-AI Role-Taking", Kalkhoff, W., Thye, S.R. and Lawler, E.J. (Ed.) Advances in Group Processes (Advances in Group Processes, Vol. 39), Emerald Publishing Limited, Leeds, pp. 1-22. https://doi.org/10.1108/S0882-614520220000039001

Publisher

:

Emerald Publishing Limited