BEYOND TRUST: : Operationalizing human-centered XAI and AI design principles to drive user experience and usability of a job recommender system
Year: 2025
Type: dissertation
Abstract: As Artificial Intelligence (AI) becomes more and more common in our everyday lives, it’s important to design explanations that everyone can understand, not just expert users. Most Explainable Artificial Intelligence (XAI) today is developer-centric, focused on exposing technical details of models and data, without considering the end-user’s goals or level of expertise. This study looks at how XAI design principles can improve the Swedish employment service’s (Arbetsförmedlingen) AI job recommendation tool, “Upptäck andra yrken” to help users better understand job suggestions and have an improved overall user experience. The study utilizes a Research through Design (RtD) methodology to guide the thesis. By conducting 6 user tests consisting of prototypes, semi-structured interviews, and a User Experience Questionnaire (UEQ) the study aims to answer two research questions: (1) How can specific Explainable AI (XAI) principles be integrated into a job recommender system to enhance users' understanding of recommendations?, and (2) To what extent does the inclusion of XAI principles in a job recommender system improve user experience? To analyze the findings of the study a thematic analysis was conducted utilizing affinity diagrams. Results show that features supporting comparisons, self-explanations, and context awareness help users understand recommendations better and improve their overall user experience. Ultimately, this study offers initial insights into how XAI design principles can be operationalized in AI-driven job recommendation context and serves as a valuable contribution for future research investigating the relationship between explainability and user experience.
Authors: Alaniemi, Peter, Bergman, Elias
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
SDG: Reduced inequalities
Open Access status: green
OpenAlex ID: W7036375500
