Impacts of AI ‐Generated Learning Videos on Learning Outcomes: The Mediating Roles of Perceived Desirability and Perceived Feasibility

Background In recent educational research and practice, there has been growing attention to the use and impact of generative AI technologies, with most studies focusing on AI‐generated textual content. However, relatively little is known about AI‐generated learning videos, despite their potential to significantly influence learners' experiences and outcomes. Objectives Drawing on situational interest theory and construal level theory, this study investigates how three different types of AI‐generated learning videos influence learning outcomes through the mediating roles of perceived desirability and perceived feasibility. Three AI‐generated video types that represent ascending levels of learner involvement—situated videos, multiple‐choice interactive situated videos, and AI–learner co‐created videos—are examined. Methods A between‐subjects experiment was conducted with 221 valid responses across four conditions: three AI‐generated learning video types and a traditional learning video control condition. ANOVA and partial least squares structural equation modelling (PLS‐SEM) were employed to validate the effects. Results and Conclusions Results show that all three types of AI‐generated videos significantly enhance learning outcomes—learning enjoyment, learning performance, and continuous learning intention—compared to traditional learning videos, with perceived feasibility mediating the effects on all three outcomes and perceived desirability mediating the effects on learning enjoyment and continuous learning intention. This study extends the theoretical boundaries of AI in learning research and provides actionable insights for teachers in instructional design.