The role of AI competencies: How do AI knowledge, skills, and attitudes shape users’ understanding and evaluation of explainable news recommendations?

This study investigates how explanations in AI-driven news recommender systems (RS) influence user understanding, trust, and attitudes toward the RS, and whether these effects depend on individual differences in AI competence. An online experiment ( N = 754) was conducted in the Netherlands with a no-explanation condition as baseline. Results show that joint explanations significantly improved users’ objective understanding of the RS but had no consistent direct effects on perceived understanding, trust, or attitudes toward the RS. Exploratory structural equation modeling revealed an indirect pathway in which explanations enhanced objective understanding, which increased perceived understanding, subsequently fostering greater trust and more positive attitudes. Moderation analyses showed limited evidence for AI competence conditioning explanation effects, with small benefits concentrated among users with higher AI knowledge. Together, these findings suggest that explanations in news RS are not universally effective: their effects appear to operate primarily through cognitive mechanisms, while competence-related differences remain limited and selective.