Would You Trust an AI Chatbot? Public Attitudes Toward AI-Mediated Support in Sensitive Contexts

Abstract

AI-driven chatbots are increasingly deployed in domains that require emotional sensitivity, confidentiality and responsible data practices. Yet little is known about how the public negotiates trust, privacy and perceived usefulness when interacting with AI systems in sensitive support contexts. This study investigates these dynamics through a mixed-methods examination of attitudes toward AI-mediated assistance, using cyberstalking as an illustrative context where emotional vulnerability and ethical concerns are heightened. A survey (n = 53) measured attitudes across the constructs of trust, privacy and usefulness, complemented by semi-structured interviews that explored users’ expectations of empathy, emotional safety and organisational responsibility. Quantitative results show strong positive associations between trust and usefulness, while privacy concerns significantly diminish both. Qualitative findings highlight two additional themes — perceived emotional authenticity and ethical alignment — which shape how individuals evaluate the appropriateness of AI-based support. Integrating these insights, the paper presents a synthesising conceptual representation showing how trust, privacy and usefulness jointly influence willingness to rely on AI chatbots in sensitive situations, situating the findings within established information-systems constructs. The study contributes to the IS discourse by demonstrating how public attitudes toward AI are shaped not only by functionality but by expectations of care, transparency and ethical stewardship. Implications for the design and governance of AI support systems are discussed.

This document is currently not available here.

Share

COinS