IIM-I Study: Trust, Not Privacy Settings, Makes Users Open Up To AI

Indore (Madhya Pradesh): As organisations increasingly deploy AI-powered chatbots and virtual assistants, a new study by the Indian Institute of Management (IIM) Indore has found that trust and perceived value play a much greater role in encouraging users to share information with AI than traditional privacy settings.

Published in the International Journal of Information Management, the study, co-authored by IIM Indore faculty member Prof Saurabh Kumar, found that human-like conversational AI builds stronger trust among users, making them more willing to disclose information.

Privacy controls alone, however, have a relatively limited influence on disclosure behaviour.

With AI becoming integral to customer service, healthcare, e-commerce and digital platforms, the research addresses a key challenge for organisations: how to encourage user engagement while protecting privacy.

The study examined how individuals decide whether to share personal information with Large Language Model (LLM)-powered conversational AI systems such as ChatGPT and similar virtual assistants.

Researchers surveyed 230 AI-aware participants from the US and UK and analysed the relationship between privacy concerns, perceived risks, trust and human-like AI interactions.

Drawing on Communication Privacy Management Theory and Privacy Calculus Theory, the researchers assessed how privacy controls, privacy-by-design features, privacy nudges and anthropomorphic characteristics influence users' willingness to share information with generative AI systems.

One of the study's key findings was that users are more likely to disclose information when they perceive clear benefits from interacting with AI and trust the system.

Human-like communication emerged as a major factor in building that trust, significantly enhancing perceptions of usefulness and value.

The study also found that privacy-by-design features and privacy nudges strengthen trust and improve perceptions of privacy protection.

However, conventional privacy settings alone have only a limited impact on disclosure behaviour.

The researchers recommend that organisations move beyond compliance-focused privacy policies and design AI systems that actively cultivate trust through transparency and responsible design.

The study concludes that the future of human-AI interaction will depend on balancing personalisation with privacy.