This study examines the use of Google Gemini Veo 3 for generating instructional videos in blended learning environments supported by a Learning Management System (LMS). A quasi-experimental mixed-methods approach was employed to evaluate the effects of AI-generated videos on students’ learning performance and engagement. Results show a statistically significant improvement in post-test scores with moderate effect size, as well as increased behavioral, cognitive, and emotional engagement. Qualitative findings indicate that students value the videos for their clarity, visual support, short length, accessibility features, and alignment with assessments, while also highlighting the importance of transparency in AI use. The findings suggest that AI-generated videos can enhance blended learning when integrated with sound instructional design and ethical practices.