ICLS Hybrid Symposium | iHY-5_AI-EMBEDDED | Evaluating and assessing learning in AI-embedded environments: perspectives across learning scientists

This hybrid poster symposium brings together eight studies that examine how evaluation can serve as a generative process for designing, enacting, and refining AI-supported learning environments. Moving beyond questions of technical accuracy or efficiency, these contributions frame evaluation as a relational and ethical practice that shapes how educators, learners, and technologies collaborate. Across diverse contexts—from middle school classrooms and tutoring programs to teacher coaching and high school psychology—researchers employ mixed methods, including discourse analytics, qualitative video analysis, and practical measures of student experience, to understand AI’s impact on collaboration, belonging, and cognition. A shared emphasis on human-centered design and co-creation underscores the importance of positioning teachers and students as partners in shaping AI systems that amplify rather than automate human intelligence. Together, the studies highlight how rigorous, equity-oriented evaluation can illuminate both the affordances and risks of educational AI.