This paper explores the integration of Generative Artificial Intelligence (AI) in education, focusing on the design and evaluation of an AI Educational Video Assistant that leverages the Cognitive Theory of Multimedia Learning (CTML). The assistant, designed to enhance multimodal learning experiences, consists of three modules: Transcription, Engagement, and Reinforcement. These modules integrate Automatic Speech Recognition (ASR) using OpenAI’s Whisper and Google’s Large Language Model (LLM) Bard. The study aims to develop and evaluate the AI assistant's effectiveness in improving learning experiences, combining human expert evaluations with automatic metrics. The research highlights the potential of AI models like Google's Bard and OpenAI's Whisper in enhancing educational content. The Transcription module uses ASR to transcribe video content, while the Engagement module employs LLMs to generate contextually relevant responses to user queries. The Reinforcement module provides additional information and concept maps to deepen understanding. The evaluation includes a structured questionnaire from nine educational experts and automatic metrics such as Content Distinctiveness and Readability scores. The results suggest positive impacts on learning experience, engagement, content organization, clarity, and usability. The study concludes that the AI Educational Video Assistant has the potential to revolutionize educational design by providing personalized and engaging learning experiences.This paper explores the integration of Generative Artificial Intelligence (AI) in education, focusing on the design and evaluation of an AI Educational Video Assistant that leverages the Cognitive Theory of Multimedia Learning (CTML). The assistant, designed to enhance multimodal learning experiences, consists of three modules: Transcription, Engagement, and Reinforcement. These modules integrate Automatic Speech Recognition (ASR) using OpenAI’s Whisper and Google’s Large Language Model (LLM) Bard. The study aims to develop and evaluate the AI assistant's effectiveness in improving learning experiences, combining human expert evaluations with automatic metrics. The research highlights the potential of AI models like Google's Bard and OpenAI's Whisper in enhancing educational content. The Transcription module uses ASR to transcribe video content, while the Engagement module employs LLMs to generate contextually relevant responses to user queries. The Reinforcement module provides additional information and concept maps to deepen understanding. The evaluation includes a structured questionnaire from nine educational experts and automatic metrics such as Content Distinctiveness and Readability scores. The results suggest positive impacts on learning experience, engagement, content organization, clarity, and usability. The study concludes that the AI Educational Video Assistant has the potential to revolutionize educational design by providing personalized and engaging learning experiences.