Differential Effects of Process- and Purpose-Based Algorithmic Transparency on Trust Formation and Maintenance in Financial Decision-Making

Algorithmic transparency is crucial for building user trust in artificial intelligence (AI) systems. However, few studies have examined how different forms of transparency affect trust across distinct stages of interaction. This research examines how process-based and purpose-based transparency differentially affect trust formation and maintenance in algorithmic financial advisors. Study 1 employed a within-subjects design and revealed that process transparency enhanced perceived competence, whereas purpose transparency increased perceived benevolence and usage intentions. Study 2 employed a 2 (Expectation: fulfilled vs. unfulfilled) × 2 (Transparency: process-based vs. purpose-based) mixed design and found that fulfilled expectations strengthened competence beliefs and investment allocations. Process transparency further boosted high-risk investments, signaling instrumental confidence, while purpose transparency fostered more conservative reallocations and sustained benevolence-based trust. Performance outcomes overrode transparency effects post-feedback, as no significant interaction emerged. For designing AI advisors, process transparency is key for initial trust, while purpose transparency sustains engagement by aligning algorithmic goals with user values. Managing expectations is crucial for long-term trust.