Developing XR UX heuristic and applying them via AI models : Focusing on XR game content

This study posits that the stagnation in the XR game market’s growth stems chiefly from the absence of UX heuristics. Accordingly, we formalize thirteen UX heuristic principles that reflect the distinctive properties of XR environments. We selected thirteen major scholarly articles published since 2019 and conducted open coding, integrating the codes into higher order categories. We then applied these categories to seven representative XR games to assess UX elements by criterion. To enhance objectivity and efficiency in the evaluation process, we built a retrieval augmented generation (RAG) based AI UX evaluation model trained on 1,515 case instances. Given in-game images and contextual descriptions as input the model analyzes UX elements against the thirteen principles and proposes improvements. Analyzing the model’s effectiveness separately for a gamer group and an HCI researcher group, we found that the thirteen heuristic principles served as a common evaluation framework, improving inter rater agreement on terminology and categories. By addressing prior limitations of heuristic analysis related to standardization and consistency, the proposed framework and model can enhance the practical utility of UX evaluation for XR games.