The incorporation of emerging technologies in education has enabled the exploration of new methodologies to enhance teaching and learning. In this project, video micro-lessons on differential and integral calculus were developed using virtual avatars created with the Heygen software. The videos present key course content in a dynamic and accessible format, initially using generic avatars, with the option of personalization by digitizing the image of any instructor. The scripts for the micro-lessons were initially prepared by subject-matter experts. However, to save time and improve the quality of the writing, the scripts were reviewed and refined using ChatGPT, which helped adjust the language and better structure the content. This process resulted in audiovisual materials that were clear, concise, and aligned with the course objectives. The design of the micro-lessons focused on selecting small but fundamental topics to facilitate understanding of the main ideas in differential and integral calculus. By breaking the content into manageable segments, progressive learning was promoted, and cognitive overload for students was avoided.

The work was implemented using the flipped learning methodology. Students watched the videos before class and completed a reinforcement quiz designed to consolidate the micro-lessons. Once in the classroom, the learning was complemented with general or more complex cases than those presented in the quiz, allowing class time to be better utilized for practical and collaborative activities. To measure the impact of the micro-lessons, a study was conducted using pretest and posttest evaluations. These assessments compared the students' performance before and after the implementation of the videos. The data revealed a significant positive impact on the students’ learning levels, demonstrating a notable improvement in their understanding of the topics covered.

Keywords: Educational Innovation, Higher Education, Generative IA, Flipped Learning.