UXexcelerateAI: A Human-Centred Framework for Evaluating Trust and Ethics in Human–Robot Interaction

This paper presents “UXexcelerateAI”, a novel evaluation framework for Human–Robot Interaction (HRI) developed through a sequence of empirical studies using participatory design and mixed-method evaluation. Across these studies, we identified systematic gaps in existing HRI evaluation practices, particularly the limited ability of established instruments such as the Godspeed Questionnaire Series (GQS) to capture relational, ethical, and interpretability-related aspects of user experience. To address these limitations, UXexcelerateAI extends GQS with six additional dimensions: perceived trust, consciousness, emotional intelligence, ethical considerations, interpretability, and user acceptance.The framework was psychometrically piloted and subsequently applied in a proof-of-concept evaluation of a GPT-powered RoboThespian Digital Twin companion designed to support adults with ADHD. Results demonstrate that, when used alongside GQS, UXexcelerateAI provides richer and more multidimensional insights into user perceptions than GQS alone, particularly in contexts involving AI-driven and support-oriented interactions. This paper contributes (1) a novel empirically grounded, extensible evaluation framework for assessing complex human–robot interactions, and (2) an applied demonstration of evaluating GPT-enabled humanoid companions in a neurodiverse support context. Future work will focus on large-scale validation, cross-cultural deployment, and longitudinal studies to establish UXexcelerateAI as a domain-agnostic framework for next-generation HRI evaluation.