Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Silver-Tongued and Sundry: Exploring Intersectional Pronouns with ChatGPT

Takao Fujii, Katie Seaborn, Madeleine Steeds

TL;DR

The paper investigates whether Japanese first-person pronouns can trigger perceptions of intersectional social identities in ChatGPT, using a mixed-method online experiment across two Japanese regions. Ten pronoun sets map onto gender, age, region, and formality, revealing systematic gender cues and nuanced region-related effects. While pronouns effectively cue simple and intersectional identities, urban standardization can dampen regional distinctions, suggesting careful design of personas for LLM-based agents. The study advances language-based methods for culturally sensitive persona development and highlights avenues for cross-linguistic and ethical research on pronoun use in AI systems.

Abstract

ChatGPT is a conversational agent built on a large language model. Trained on a significant portion of human output, ChatGPT can mimic people to a degree. As such, we need to consider what social identities ChatGPT simulates (or can be designed to simulate). In this study, we explored the case of identity simulation through Japanese first-person pronouns, which are tightly connected to social identities in intersectional ways, i.e., intersectional pronouns. We conducted a controlled online experiment where people from two regions in Japan (Kanto and Kinki) witnessed interactions with ChatGPT using ten sets of first-person pronouns. We discovered that pronouns alone can evoke perceptions of social identities in ChatGPT at the intersections of gender, age, region, and formality, with caveats. This work highlights the importance of pronoun use for social identity simulation, provides a language-based methodology for culturally-sensitive persona development, and advances the potential of intersectional identities in intelligent agents.

Silver-Tongued and Sundry: Exploring Intersectional Pronouns with ChatGPT

TL;DR

The paper investigates whether Japanese first-person pronouns can trigger perceptions of intersectional social identities in ChatGPT, using a mixed-method online experiment across two Japanese regions. Ten pronoun sets map onto gender, age, region, and formality, revealing systematic gender cues and nuanced region-related effects. While pronouns effectively cue simple and intersectional identities, urban standardization can dampen regional distinctions, suggesting careful design of personas for LLM-based agents. The study advances language-based methods for culturally sensitive persona development and highlights avenues for cross-linguistic and ethical research on pronoun use in AI systems.

Abstract

ChatGPT is a conversational agent built on a large language model. Trained on a significant portion of human output, ChatGPT can mimic people to a degree. As such, we need to consider what social identities ChatGPT simulates (or can be designed to simulate). In this study, we explored the case of identity simulation through Japanese first-person pronouns, which are tightly connected to social identities in intersectional ways, i.e., intersectional pronouns. We conducted a controlled online experiment where people from two regions in Japan (Kanto and Kinki) witnessed interactions with ChatGPT using ten sets of first-person pronouns. We discovered that pronouns alone can evoke perceptions of social identities in ChatGPT at the intersections of gender, age, region, and formality, with caveats. This work highlights the importance of pronoun use for social identity simulation, provides a language-based methodology for culturally-sensitive persona development, and advances the potential of intersectional identities in intelligent agents.
Paper Structure (33 sections, 1 equation, 6 figures, 6 tables)

This paper contains 33 sections, 1 equation, 6 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Background and Theory
  3. Intersectionality and Identity in HCI
  4. Intersectionality and Identity in AI and Intelligent Agents
  5. Identity and Japanese First-person Pronouns
  6. Theoretical Model: Mapping First-Person Pronouns to Factors of Social Identity, Intersectionally
  7. Methods
  8. Participants
  9. Procedure
  10. Materials
  11. Instruments and Measures
  12. Genderedness
  13. Agedness
  14. Similarity-to-Self
  15. Impressions of the Persona
  16. ...and 18 more sections

Figures (6)

  • Figure 1: Pronoun map with the ten types of Japanese first-person pronouns intersectionally linked to social identities in Japan. Blue dashed lines: Formality. Green solid lines: Gender. Pink dashed lines with a dot: Region. Gold dotted lines: Age.
  • Figure 2: Example with all five ChatGPT prompts.
  • Figure 3: Example of a conversation.
  • Figure 4: The gender categorization of each pronoun type as a percentage (N=201).
  • Figure 5: Age categorizations for each pronoun type as a percentage (N=201).
  • ...and 1 more figures