Checklist: RE4HCAI Framework

This research investigates how users perceive AI-based recommender systems across different usage contexts through a human-centered lens. Using a qualitative and exploratory approach, two representative systems from distinct domains were analyzed: Spotify (entertainment, hedonic context) and Duolingo (education, evaluative context). The methodology integrated three complementary stages: (i) application of the RE4HCAI framework to assess human-centered AI requirements; (ii) development of empathy maps based on representative personas; and (iii) interpretive triangulation between technical requirements and experiential insights. Findings reveal that AI perception is not directly tied to technical transparency, but rather to the emotional impact of algorithmic decisions on user experience.This spreadsheet presents the structured results of the RE4HCAI checklist-based evaluation. The dataset is organized in a tabular format containing: (i) the checklist item, (ii) its corresponding requirement description, and (iii) a categorical assessment indicating whether the requirement was identified in the system ("Yes"), not identified ("No"), or partially implemented ("Partially"). The evaluation covered six dimensions: User Needs, Model Needs, Data Needs, Feedback & User Control, Explainability & Trust, and Errors & Failures. This dataset enables reproducibility of the evaluation process and may support future research on AI transparency, explainability, and human-centered system design.