Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Longitudinal Randomized Control Study of Companion Chatbot Use: Anthropomorphism and Its Mediating Role on Social Impacts

Rose E. Guingrich, Michael S. A. Graziano

TL;DR

This longitudinal randomized study tests whether daily companion chatbot use harms or helps social health, loneliness, or relationships. It uses random assignment to a companion chatbot vs non-social word games over 21 days with four surveys and two interviews, totaling 183 participants. The main finding is no overall social health impact, but a mediated pathway shows that individuals with higher social-connection desire anthropomorphize the chatbot more, and greater anthropomorphism predicts larger social impacts on human relationships, supporting a mediation model. These results highlight the role of social motivation and mind attribution in translating AI interactions into human social outcomes, with implications for AI design and public concerns about AI-mediated sociality.

Abstract

Many Large Language Model (LLM) chatbots are designed and used for companionship, and people have reported forming friendships, mentorships, and romantic partnerships with them. Concerns that companion chatbots may harm or replace real human relationships have been raised, but whether and how these social consequences occur remains unclear. In the present longitudinal study ($N = 183$), participants were randomly assigned to a chatbot condition (text chat with a companion chatbot) or to a control condition (text-based word games) for 10 minutes a day for 21 days. Participants also completed four surveys during the 21 days and engaged in audio recorded interviews on day 1 and 21. Overall, social health and relationships were not significantly impacted by companion chatbot interactions across 21 days of use. However, a detailed analysis showed a different story. People who had a higher desire to socially connect also tended to anthropomorphize the chatbot more, attributing humanlike properties to it; and those who anthropomorphized the chatbot more also reported that talking to the chatbot had a greater impact on their social interactions and relationships with family and friends. Via a mediation analysis, our results suggest a key mechanism at work: the impact of human-AI interaction on human-human social outcomes is mediated by the extent to which people anthropomorphize the AI agent, which is in turn motivated by a desire to socially connect. In a world where the desire to socially connect is on the rise, this finding may be cause for concern.

A Longitudinal Randomized Control Study of Companion Chatbot Use: Anthropomorphism and Its Mediating Role on Social Impacts

TL;DR

This longitudinal randomized study tests whether daily companion chatbot use harms or helps social health, loneliness, or relationships. It uses random assignment to a companion chatbot vs non-social word games over 21 days with four surveys and two interviews, totaling 183 participants. The main finding is no overall social health impact, but a mediated pathway shows that individuals with higher social-connection desire anthropomorphize the chatbot more, and greater anthropomorphism predicts larger social impacts on human relationships, supporting a mediation model. These results highlight the role of social motivation and mind attribution in translating AI interactions into human social outcomes, with implications for AI design and public concerns about AI-mediated sociality.

Abstract

Many Large Language Model (LLM) chatbots are designed and used for companionship, and people have reported forming friendships, mentorships, and romantic partnerships with them. Concerns that companion chatbots may harm or replace real human relationships have been raised, but whether and how these social consequences occur remains unclear. In the present longitudinal study (), participants were randomly assigned to a chatbot condition (text chat with a companion chatbot) or to a control condition (text-based word games) for 10 minutes a day for 21 days. Participants also completed four surveys during the 21 days and engaged in audio recorded interviews on day 1 and 21. Overall, social health and relationships were not significantly impacted by companion chatbot interactions across 21 days of use. However, a detailed analysis showed a different story. People who had a higher desire to socially connect also tended to anthropomorphize the chatbot more, attributing humanlike properties to it; and those who anthropomorphized the chatbot more also reported that talking to the chatbot had a greater impact on their social interactions and relationships with family and friends. Via a mediation analysis, our results suggest a key mechanism at work: the impact of human-AI interaction on human-human social outcomes is mediated by the extent to which people anthropomorphize the AI agent, which is in turn motivated by a desire to socially connect. In a world where the desire to socially connect is on the rise, this finding may be cause for concern.

Paper Structure

This paper contains 32 sections, 3 figures.

Table of Contents

  1. Introduction
  2. Theory of Mind
  3. Anthropomorphism
  4. Social Need
  5. The Present Study
  6. Methods
  7. Participants
  8. Study Design
  9. Scales
  10. Sequence of Events
  11. Results
  12. Participant Retention
  13. Participant Engagement
  14. Assumptions Testing
  15. Trends Over Time
  16. ...and 17 more sections

Figures (3)

  • Figure 1: Relationships between desire to socially connect and anthropomorphism for the chatbot condition. Desire to socially connect on day 1 is plotted on the x-axes, and anthropomorphism scores are plotted on the y-axes. A. Relationship between desire to socially connect on day 1 and anthropomorphism of the chatbot averaged over days 7, 14, and 21. B. Relationship between desire to socially connect on day 1 and anthropomorphism scores on days 7, 14, and 21, plotted separately. The data for day 7 is plotted with squares, for day 14 is plotted with circles, and for day 21 is plotted with triangles.
  • Figure 2: Relationships between anthropomorphism and magnitude of social impact for the chatbot condition. Anthropomorphism is plotted on the x-axes, and the magnitude of social impact is plotted on the y-axes. A. Relationship between anthropomorphism of the chatbot averaged over time and magnitude of social impact averaged over time. B. Relationship between anthropomorphism scores and respective magnitude of social impact on days 7, 14, and 21. The data for day 7 is plotted with squares, for day 14 is plotted with circles, and for day 21 is plotted with triangles.
  • Figure 3: Mediation model for desire to socially connect, magnitude of social impact, and anthropomorphism. Paths between variables designated with letters $a$, $b$, and $c$. Path estimates and significance are indicated (* = $p < 0.01$, ** = $p < 0.001$, ns = non-significant).