The Best Paper Award of the 13th International Conference on HCI in Business, Government and Organizations
has been conferred to
Ancuta Margondai, Sara Willox, Anamaria Acevedo Diaz, Julie Rader, Soraya Hani, Valentina Ezcurra
and Mustapha Mouloua (University of Central Florida, USA)
for the paper entitled
"The Transparency Dilemma: Reconciling Trust Calibration with Perceived Autonomy Through Personality-Adaptive AI Systems"
Ancuta Margondai
(presenter)
Best Paper Award for the 13th International Conference on HCI in Business, Government and Organizations, in the context of HCI International 2026, Montreal, Canada, 26 - 31 July 2026
Certificate for Best Paper Award of the 13th International Conference on HCI in Business, Government and Organizations presented in the context of HCI International 2026, Montreal, Canada, 26 - 31 July 2026
Paper Abstract
Artificial intelligence systems increasingly support high-stakes organizational decision-making, yet regulatory frameworks mandating transparency assume detailed explanations universally improve human-AI collaboration. Evidence suggests a paradox: computational models demonstrate that transparency enhances trust calibration–the alignment between user trust and actual AI reliability–while empirical studies reveal the same features reduce perceived autonomy and can trigger cognitive overload. This contradiction presents a critical challenge for AI design and policy. The present research integrated computational modeling with empirical data across three studies to demonstrate that personality traits systematically moderate transparency’s effects. Study 1 employed agent-based simulation (N = 24,000 agents, 24 million decisions) modeling human-AI collaboration across four transparency levels and three reliability conditions. Study 2 collected empirical data from 557 organizational decision-makers using a factorial design crossing transparency conditions with Big Five personality traits. Study 3 extended the simulation with personality parameters derived from empirical regression analyses, enabling direct validation. The personality-enhanced simulation reproduced empirical findings with exceptional fidelity (interaction coefficient differences = 0.004). Results revealed that transparency creates dissociable effects on trust calibration and autonomy perception: high transparency improved trust calibration universally (), but autonomy effects were personality-dependent. Individuals high in openness exhibited autonomy gains under transparency (), while individuals high in extraversion experienced amplified autonomy loss (). Population-level analysis indicated universal high transparency creates optimal outcomes for only 18% of users. Findings challenge one-size-fits-all transparency mandates and provide an empirically validated framework for personality-adaptive AI systems.
The full paper is available through SpringerLink, provided that you have proper access rights.