Frameworks for AI in Education: Methods and Problems
Speaker: Prof. Zeev (Vladimir) Volkovich
Dr. Renata Avros
Abstract: This presentation explores the growing impact of generative AI models on education and the new challenges they pose for teaching and learning. It examines the shift from information-retrieval-based approaches to generative intelligence, emphasizing the importance of prioritizing inquiry, reasoning, and understanding over the mere production of final answers. The talk identifies common risks associated with AI use in learning, including shortcut-seeking behavior, over-reliance on automated solutions, and superficial understanding. To address these challenges, the presentation discusses practical pedagogical strategies based on the principle that AI should amplify, rather than replace, human thinking. These include the “Explain-It-Back” method, structured weekly AI-enhanced learning routines, and clear guidelines for ethical AI use. A Generative Deep Learning course framework and a neural style transfer case study further support the discussion. In addition, the presentation includes a discussion of a student project developed within a cryptology course, illustrating how AI techniques can be applied to cryptographic analysis while maintaining rigorous validation, critical thinking, and academic integrity.