Understanding user trust in AI-based recommendation tool for enhancing social media profiles

Abstract

As AI-driven recommendation tools become more prevalent in digital interactions, user trust has emerged as a critical factor influencing their adoption and effectiveness. This thesis examines user trust and adoption of TrustifyAI, an AI-based recommendation tool designed to enhance the credibility of Twitter profiles, using the Technology Acceptance Model 3 (TAM3) as the guiding framework. Addressing three research questions on how perceived ease of use (PEOU), perceived usefulness (PU), and behavioral intention (BI) influence trust in AI-generated recommendations, the study employed a mixed-methods approach, data from 16 participants were analyzed. Quantitative results indicate high usability and perceived relevance of the tool’s suggestions, with participants rating the interface as intuitive and the feedback as actionable. Trust was measured not only as an emergent outcome but as a distinct construct, reflected in high scores on trust-related items and supported by qualitative insights. Qualitative responses reveal that the tool encouraged self-reflection and improved awareness of digital credibility cues. Participants also expressed a desire for greater personalization, transparency, and platform flexibility. The findings affirm the applicability of TAM3 in evaluating AI-driven digital tools and highlight design priorities—such as clarity, ethical operation, and user empowerment—that are essential for fostering AI systems for digital identity management.