Digital Literacy as a Predictor of AI Literacy: A Gender Paradox and a Knowledge–Belief Gap

This study investigates the relationship between university students’ digital literacy and artificial intelligence (AI) literacy. Participants were 770 undergraduate students (227 male, 543 female; Mage = 21.27, SD = 2.75) recruited through online convenience and snowball sampling, with recruitment centered at Karamanoğlu Mehmetbey University (Karaman, Türkiye), representing more than 120 academic programs. Data were collected with the Turkish forms of Ng’s (2012) Digital Literacy Scale (α = .931) and Wang et al.’s (2022) AI Literacy Scale (α = .803). Three findings stand out. First, digital literacy and AI literacy were strongly and positively related (r = .596, p < .001). A three-step hierarchical regression analysis identified digital literacy as the only significant predictor of AI literacy (β = .599, p < .001), explaining 35.5% of the variance (R² = .355, F(1, 768) = 422.57, p < .001); adding gender and age (ΔR² = .002) and internet use frequency (ΔR² = .000) did not significantly improve the model. Second, a small gender difference favoring men emerged in digital literacy (Cohen’s d = 0.201) but not in AI literacy (d = 0.031), a pattern we describe as a gender paradox. Third, although 85.1% of participants believed in AI’s future impact, only 22.7% rated their own AI knowledge as sufficient—a knowledge–belief gap. Self-rated AI knowledge, but not belief, differed significantly by gender (χ² = 11.05, p = .004; Cramér’s V = .120). Implications for digital and AI literacy education in higher education are discussed.