Thinking Through AI: Advancing Cognitive and Collaborative Research for AI in Education

This paper introduces the "Thinking Through AI" framework, an innovative methodological approach designed to explore and enhance cognitive and collaborative processes in educational settings mediated by artificial intelligence (AI). Based in a synthesis of cognitive labs, think-aloud protocols, over-the-shoulder observations, and cyber-social research method, the framework addresses key methodological challenges in studying AI's role in education. The framework emphasizes three core dimensions: (1) students’ cognitive engagement with AI, fostering their ability to achieve learning objectives while simultaneously improving AI systems through interactive feedback; (2) educators’ pedagogical strategies, integrating rubrics and knowledge bases to design AI tools tailored to student needs; and (3) the concurrent design of AI tools and learning experiences, informed by think-aloud and cognitive labs as part of Learning Experience Research (LXR). To support and exemplify this framework, the study integrates data from the implementation of the CGScholar AI Helper at a rural Alaska school involving 30 students. This implementation provides empirical insights into how AI-enabled tools can enhance writing instruction in middle school classrooms, emphasizing the practical application and refinement of the framework. This approach redefines the intersection of user experience and educational design in the AI era, positioning educators, learners, and developers as co-creators in the design process. By enabling educators to actively contribute to AI tool creation, the framework bridges pedagogical practices with AI advancements, ensuring tools align with diverse learning objectives and teaching methodologies. This paper highlights the transformative potential of collaborative methodologies, offering a practical and theoretical contribution to advancing AI's role in education.