Automating Assessments in Python and Data Science Using Cognitive-Aware Generative AI

Generative AI is increasingly being integrated into education, offering solutions to automate various tasks. One of the most time-consuming responsibilities for educators is creating quizzes and assessments, which are essential for evaluating student learning. While AI can help automate this process, its effectiveness depends on how well the generated questions support foundational learning rather than just producing random assessments. Many AI-generated quizzes lack a structured approach to assessing beginner and novice learners. Effective assessments should focus on foundational cognitive processes, such as remembering key concepts, and understanding their meaning, before progressing to higher-order thinking skills. Bloom’s Taxonomy provides a useful framework, where the initial cognitive levels— Remember and Understand —are essential for building a strong learning foundation. This paper presents a cognitive-aware generative AI framework that creates quizzes and assessments aligned with the remember and understand levels of Bloom’s Taxonomy . Using few-shot prompting and adaptive question generation, the system produces beginner-friendly multiple-choice, true/false, and fill-in-the-blank questions that emphasize clarity and conceptual comprehension. The effectiveness of this approach is evaluated through educator reviews to determine their quality, relevance, and cognitive level. Additionally, the study will also explore how this system reduces educators’ workload while maintaining educational integrity. Our preliminary evaluation results confirm the validity and effectiveness of the proposal.