2023journal article|102 цитирований

Friend, mentor, lover: does chatbot engagement lead to psychological dependence?

Tian Xie, Iryna Pentina, Tyler Hancock

Journal of Service Management

Перейти на сайт издателя Semantic Scholar

Краткое изложение простым языком

This mixed-method study explores whether deep engagement with AI social chatbots can lead to psychological dependence. Users assign various relationship roles to their chatbot companions, from friends to romantic partners. While chatbot interactions can provide genuine emotional support, the research identifies risk factors for unhealthy dependence, particularly among lonely or socially anxious users. The findings raise important questions about the mental health implications of increasingly sophisticated AI companions.

Аннотация

PurposeThe purpose of this study is to explore customer-artificial intelligence (AI) service technology engagement and relationship development drivers, as well as potential negative consequences in the context of social chatbots.Design/methodology/approachA sequential mixed-method approach combined exploratory qualitative and confirmatory quantitative analyses. A conceptual model developed from Study 1 qualitative content analysis of in-depth interviews with active users of the AI social chatbot Replika was tested in Study 2 by analyzing survey data obtained from current Replika users.FindingsLoneliness, trust and chatbot personification drive consumer engagement with social chatbots, which fosters relationship development and has the potential to cause chatbot psychological dependence. Attachment to a social chatbot intensifies the positive role of engagement in relationship development with the chatbot.Originality/valueThis study was the first to combine qualitative and quantitative approaches to explore drivers, boundary conditions and consequences of relationship and dependence formation with social chatbots. The authors proposed and empirically tested a novel theoretical model that revealed an engagement-based mechanism of relationship and dependence formation with social chatbots.

Key Findings

Sustained engagement with AI social chatbots can lead to psychological dependence in some users
Users develop diverse relationship roles with chatbots including friend, mentor, and romantic partner
Loneliness and social anxiety are significant predictors of deeper chatbot engagement and dependence risk
The perceived empathy and responsiveness of AI chatbots drive continued engagement and relationship deepening
A sequential mixed-method approach revealed both positive well-being effects and potential dependency concerns