Designing Technologies for Critical and Active Learning in the Age of AI

Dr. Anjali Singh, University of Texas at Austin.

Abstract: As students increasingly turn to AI for learning support, it becomes critical to develop methods for teaching them to use AI in ways that augment rather than undermine their learning. In this talk, I will present my research on designing, developing, and evaluating learning technologies that embrace AI while demanding active cognitive engagement. First, I will introduce MetaCues, an interactive system that analyzes learning behavior and AI interactions in real time to deliver cues that promote critical thinking and active learning. I will then present experimental findings on personalized learning experiences for data science learners, demonstrating how independent thinking before interacting with AI improves both engagement and the quality of students’ work. Finally, I will outline my future research agenda focused on: (i) designing technologies that empower teachers to create comprehensive student-AI collaborative learning experiences, that build disciplinary knowledge and skills aligned with the evolving nature of work, and (ii) advancing methods to assess learning by capturing not only outcomes, but also the very process of learning.

Bio: Dr. Anjali Singh is a postdoctoral fellow at the University of Texas at Austin, conducting research at the intersection of Learning Sciences and Human-AI Interaction. Her research focuses on the design and study of interactive learning technologies that promote learners’ critical thinking and agency. She holds a PhD in Information from the University of Michigan, and her work has been recognized with multiple best paper awards.

Effort-Sensitive AI for Learning: Measuring and Shaping Persistence

Conrad Borchers, CMU.

Abstract: Highly capable learning tools, including LLMs, risk inducing overreliance that undermines student effort. What if we designed educational technology to improve effort, not only achievement and metacognition? This talk outlines a research agenda for an effort-sensitive AI lab that measures, predicts, and scaffolds persistence from learning process data across K-12, vocational, and higher education. Analytics of when, how, and why students increase workload, maintain effort, or quit will open new dimensions of personalization, adaptive recommendations, and feedback that enable improved classroom management and support instructional as well as curricular design. Grounded in field studies and motivation theory, this research models persistence, procrastination, and task choice to co-design technologies that strengthen long-term self-regulation skills that predict life outcomes.

Bio: Conrad Borchers is a PhD researcher at Carnegie Mellon University’s Human-Computer Interaction Institute, advised by Vincent Aleven and Ken Koedinger. His research combines HCI and machine learning to design intelligent systems that measurably support learner persistence, effort, and self-regulation, with recent work focusing on adaptive goal setting, hybrid tutoring, and workload modeling. He holds an MSc in Social Data Science (Oxford) and a BSc in Psychology (Tübingen).

Designing Interfaces that Help People Think, Learn, and Create with AI.

Frederic Gmeiner , CMU.

Abstract: My research investigates how AI systems can be designed to actively support learning and reflective thinking in complex, open-ended creative tasks. Grounded in cognitive and learning sciences, I develop and study interaction techniques—such as reflective voice agents and adaptive graphical interfaces—that help users externalize their reasoning, explore alternatives, and critically engage with AI outputs. Building on these foundations, I envision adaptive, multimodal learning platforms that act as “thinking partners,” blending dialogue, visualization, and contextual support to enable in situ, continual learning and foster deeper cognitive engagement and creativity across fields from engineering to education. This vision aims to shape the future of learning technologies so that AI truly amplifies—rather than merely automates—human intelligence.

Bio: Frederic Gmeiner is a PhD candidate in Human-Computer Interaction at Carnegie Mellon University. He studies how to design AI interactions that foster reflective human thinking, creativity, and problem-solving in complex open-ended tasks. His work has been published in premier HCI venues, including a Best Paper Honorable Mention at ACM CHI, and is supported by organizations such as the National Science Foundation, Microsoft Research, Autodesk, Google, and Prolific.

Learning How to Learn in the Age of AI: Measuring, Understanding, and Supporting Self-Regulated Learning

Joyce (Jiayi) Zhang, University of Pennsylvania.

Abstract: As AI tools like ChatGPT transform how we access knowledge, the critical question becomes not what students can learn, but how they learn. This talk explores self-regulated learning (SRL; the skill that enables students to plan, monitor, and adapt their learning) in digital environments. I will share research on detecting SRL at scale, understanding how it interacts with affect, engagement, and motivation, and designing interventions that foster strategic learning. Looking ahead, I will discuss how SRL shapes whether learners use AI tools to extend their thinking or bypass meaningful learning.

Bio:Joyce (Jiayi) Zhang is a Ph.D. candidate in the Learning Science and Technology program at the University of Pennsylvania. Her research examines how to efficiently and responsibly leverage artificial intelligence to enhance student learning through data-driven approaches. She focuses on modeling how students feel, engage, and regulate their learning, with the goal of designing technologies that foster curiosity, persistence, and equitable learning opportunities for all learners.

Human-AI Partnerships for Educational Content Generation, Evaluation, and Personalization

Prof. Steven Moore, George Mason University.

Abstract. In this talk, I present a program of research that strategically integrates human intelligence with AI to co-create, evaluate, and map learning content. I first show how learnersourcing empowers students as co-creators of instructional activities, then introduce a toolkit for assessing the pedagogical validity of educational questions. Through this work, I present my cohesive vision for creating next-generation educational technology where humans and AI collaborate to continuously improve educational content, assessments, and personalization.

Bio. Steven Moore is an Assistant Professor in the Department of Information Sciences and Technology at George Mason University. His research combines expertise in learning science, computer science, and applied natural language processing to create and evaluate educational content and technology for learning at scale. Recently, his research has focused on leveraging LLMs to support instructional design, such as applying rubrics to different content types and creating simulated students. StevenJamesMoore.com

Preparing Students for an AI-Empowered Future by Modeling and Fostering 21st Century Skills

Alex Scarlatos, University of Massachusetts Amherst.

Abstract. As generative AI becomes more capable and widely adopted, it will increasingly become an integral part of daily and professional life. Education will be greatly affected: AI systems will take on a greater role to assist with lesson planning, mediate group work, and provide individualized tutoring. This shift will also redefine what students need to learn, preparing them for the future of work by placing a greater emphasis on creativity, strategic thinking, and effective collaboration with AI agents. A central challenge for educational research will be how to teach and assess these increasingly important 21st century skills, moving beyond traditional tests to capture students’ creativity, communication, and collaboration. My recent work contributes to this direction by developing methods for modeling student knowledge in dialogues with AI agents and simulating student responses in open-ended settings, which will be essential for evaluating the effectiveness of future educational systems. In a new lab, I aim to focus on cultivating these higher-order skills in open-ended learning  environments, leveraging AI-based formative assessments, student simulation, and educational systems trained to maximize student outcomes.

Bio. Alex is a Computer Science Ph.D. candidate at the University of Massachusetts Amherst, working with Dr. Andrew Lan in the machine learning for education (ML4Ed) lab. His work focuses on applying language models to tasks in educational content generation and student simulation, particularly for grade-level math. He has published research in top natural language processing conferences such as ACL, EMNLP, and NAACL, and AI for education conferences, such as AIED, LAK, and EDM; he received the best student paper award at AIED in 2024 for his work on using reinforcement learning to improve LLM-generated feedback. He has recently worked as a student researcher at Google DeepMind and an AI research intern at Cambium Assessment. He received his Master’s and Bachelor’s degrees in Computer Science from Stony Brook University, and previously worked as a software engineer at Bloomberg L.P.

Designing Multimodal Learning Analytics to Enhance Face-to-Face Collaboration

Dr. Linxuan Zhao, Monash University.

Abstract: Multimodal Learning Analytics (MMLA) is an emerging field focused on understanding and supporting complex learning activities in authentic, face-to-face settings. In this field, designing systems that effectively capture and analyse relevant data in diverse scenarios remains a significant challenge. This talk details the creation of an MMLA system for two distinct learning environments: 1) a physical healthcare simulation, where the system was used for supporting the development of learners’ teamwork skills; 2) a technical tutorial, where the system facilitated tutors in enhancing their team teaching strategy. This talk will present how this system captures multimodal data to generate actionable feedback for both learners and instructors, enhancing the learning or teaching experience. Finally, this talk will highlight key implications of this work and propose future research directions for the MMLA field.

Bio: Dr. Linxuan Zhao is a Postdoctoral Research Fellow at the Centre for Learning Analytics at Monash (CoLAM), Monash University. He specializes in multimodal learning analytics, using sensor data to understand and support students’ learning activities in authentic, face-to-face settings. His work aims to translate multimodal sensor data into actionable analytics that provide meaningful feedback to learners and instructors, ultimately facilitating a more effective learning experience.

Future-Proof’ Learning Technologies: Designing for Pedagogical and Technological Sustainability

Dr. Rania Abdelghani, Tübingen University.

Title: Abstract: As evolving learner needs and environments accelerate the obsolescence of educational technologies, building for resilience is critical. In this talk, I explore three directions for thinking about sustainable learning technologies (LTs): 1) creating generative, open-ended systems that co-evolve with users and their emergent needs; 2) architecting affordable platforms that transfer across diverse school infrastructures; and 3) developing evidence-based heuristics to measure sustainability and long-term impact of LTs.

Bio I am a postdoctoral researcher at the Hector institute, at the University of Tubingen, Germany. Before this, I earned a PhD in cognitive Science at INRIA University of Bordeaux, France and an engineering degree at INSA, University of Tunis. My work focuses on building learning technologies that foster curiosity and question-asking skills in students.

AI for Personalized Learning at Scale

Dr. Robin Schmucker, JPMorgan.

Abstract: Digital learning technologies democratize access to personalized instruction and learning materials. While tremendous progress has been made, breakthroughs in machine learning, combined with the vast data available in today’s learning platforms, offer transformative potential for education and learning science. Drawing on a series of case studies, we demonstrate (i) how reinforcement learning can enhance instructional decisions in large-scale online education, and (ii) how generative AI is shaping a new generation of more effective and scalable learning technologies. We conclude with a discussion of the broader implications of AI for education and future directions for research.

Bio: Robin Schmucker is a Senior Associate in the Machine Learning Center of Excellence at JPMorgan Chase. His research focuses on machine learning and human-AI interaction, particularly in the context of education and the future of work. He completed his PhD in the Machine Learning Department at Carnegie Mellon University, advised by Prof. Tom Mitchell. As a postdoctoral researcher, he held joint affiliations at CMU and UC Berkeley. At the CK-12 Foundation, his algorithms for student knowledge modeling and instructional decision making have enhanced learning outcomes for millions worldwide.