AI and Education Initiative’s Latest Research and Projects

PI: Leslie Dietiker, Ph.D., Associate Dean for Research and Associate Professor in Mathematics Education, Boston University Wheelock College of Education and Human Development
Co-PI: Meghan Riling, Ph.D., Assistant Professor of the Practice in Mathematics Education, Vanderbilt University Peabody College
Co-PI: Matthew Melville, Ph.D., Assistant Professor of Mathematics Education, Purdue University Fort Wayne 

Project Summary: The project will explore the ways in which ChatGPT can support teachers planning, while addressing our concerns about the lack of inquiry and student-centered features of ChatGPT-generated mathematics lesson plans. The reserachers aim to build an investigative team to explore the potential of ChatGPT for supporting the design of high school mathematics lessons. As we collectively design and refine lessons using ChatGPT utilizing GPT-4 via https://chat.openai.com, we will examine:

1. How can high school mathematics teachers use ChatGPT to plan inquiry lessons?
2. What kinds of interactions (e.g., prompt engineering, feeding the AI new information) enable high school mathematics teachers to shift lesson designs toward inquiry?
3. What support do high school mathematics teachers need when planning with ChatGPT?

PI Bio: Leslie Dietiker, Ph.D., is the associate dean for research at Wheelock College of Education and Human Development and an associate professor of mathematics education. Her research focuses attention on the curricular perspectives of mathematics teachers and their use of mathematics curriculum materials. Dietiker is the recipient of the Boston University Metcalf Award for Excellence in Teaching. Dietiker is lead author and editor of multiple mathematics textbooks for middle and high school through CPM Educational Program. She is also a fellow of the International Society for the Design and Development of Education (ISDDE). Previously, Dr. Dietiker was a mathematics teacher in San Francisco’s Phillip and Sala Burton Academic High School for 17 years.

PI: Kathy MinHye Kim, Ph.D., Clinical Assistant Professor of TESOL & Applied Linguistics and Director of the Second Language Acquisition Lab, Boston University Wheelock College of Education & Human Development
Co-PI: Dan Richard Isbell, Ph.D., Assistant Professor of Second Language Studies, University of Hawaii at Manoa
Co-PI: Dan Nickolai, Ph.D., Associate Professor of French and Director of the Language Resource Center, Saint Louis University

Project Summary: The project aims to examine theoretical questions about first language (L1) and second language (L2) proficiency and native-speaker convergence by developing a platform called Multilingual, Autoscored Elicited Imitation Test (MultiAutoEIT). This is an open-access, online tool for measuring oral language proficiency across multiple languages. By enabling automated speech recognition (ASR) and autoscoring of the responses, MultiAutoEIT offers essential tools for researchers and educators to produce controlled stimuli for diverse linguistic research and educational contexts. Additionally, it can establish benchmarks for bilingual proficiency and explore bilingual development.

PI Bio: Dr. Kathy MinHye Kim is a clinical assistant professor and director of the Language Education program at Boston University Wheelock College of Education & Human Development. Dr. Kim’s work specializes in understanding how adult learners encode and retain second language (L2) knowledge. She examines how L2 learners develop implicit and explicit knowledge and factors that mediate their development; these factors include situational features (e.g., input modality), sleep conditions, and individual differences in cognitive (e.g., intelligence, working memory capacity) and biographical (e.g., education level and reading experiences) variables.

PI: TJ McKenna, Ph.D., Director of the Center for STEM Professional Learning at Scale and Lecturer in Science education, Boston University Wheelock College of Education and Human Development
Co-PI: William Tomlinson, Ph.D., Director, Software & Application Innovation Lab (SAIL), Boston University Hariri Institute for Computing
Co-PI: Arezoo Sadeghi, Senior Software Engineer, Software & Application Innovation Lab (SAIL), Boston University Hariri Institute for Computing

Project Summary: The project aims to empower under-resourced and marginalized communities who lack necessary resources for high-quality STEM curricula. The researchers will develop a STEM EquityCoach, a chatbot specifically trained to assist classroom teachers in customizing their STEM curricula. By leveraging the capabilities of OpenAI, this innovative tool will streamline the process of creating personalized STEM lessons. The solution is designed to address the pressing need for culturally relevant and inclusive STEM education, ensuring that every student, regardless of their background, has equal access to high-quality learning experiences.

PI Bio: Thomas “TJ” McKenna, Ph.D., is a lecturer in science education at BU Wheelock College of Education & Human Development. He has spent the past 20 years as a scientist, educator (formal and informal), on-air science communicator, and faculty member focused on uncovering what works well in STEM education and how we can improve in scaling research and practice-based efforts. McKenna works at the nexus of policy, practice, and research in ways that deepen the understanding of how to make progress on the ecological problem of supporting teachers (preservice & in-service) in their efforts to offer high-quality science learning to all our students. He is also the founder and creator of Phenomena for NGSS, an educational website designed to support teachers in learning more about phenomena-based instruction.

PI:  Zachary (Zach) Rossetti, Ph.D.,Associate Professor of Special Education, Boston University Wheelock College of Education & Human Development 
Co-PI: Manuel (Manny) Ramirez, Ph.D. student in Educational Studies with a specialization in special education, Boston University Wheelock College of Education & Human Development 

Project Summary: This study will develop and pilot a protocol for using and teaching generative artificial intelligence (AI) to engage in comprehensible and culturally responsive translation work to address persistent problems of language access for culturally and linguistically diverse (CLD) families navigating the special education system. The project will employ a community-based participatory research (CBPR) design (Minkler & Wallerstein, 2011) to center the expertise of multilingual, multicultural families in developing a protocol for using and teaching AI to conduct comprehensible and culturally responsive translation and information dissemination in special education. The researchers will a) develop the protocol with an Advisory Group of five Spanish-speaking caregivers of children with disabilities, and b) pilot the protocol with a second Advisory Group of five Vietnamese caregivers of children with disabilities.

PI Bio: Zachary Rossetti, Ph.D., is an associate professor of special education at BU Wheelock College of Education & Human Development. Dr. Rossetti’s research focuses on the experiences of families with children with intellectual and developmental disabilities (IDD) by centering on participation and language access in IEP meetings, family-professional partnerships of culturally and linguistically diverse families, family advocacy, community participation, and sibling roles and relationships. His research also examines social inclusion and belonging of individuals with IDD, specifically the contexts and dynamics of friendships and how educators and families can facilitate friendship opportunities between students with and without IDD. Central to all of his research is the context of inclusive education for students with IDD.

PI:  TJ McKenna , Ph.D.,Director, Center for STEM Professional Learning at Scale Clinical Assistant Professor
Co-PI: William Tomlinson , Director, Software & Application Innovation Lab (SAIL)

Project Summary: TJ McKenna and SAIL are working together to build an AI coach to help teachers sift through our database of high quality STEM materials to choose things that are uniquely suited to their needs and their students.

PI Bio: Dr. Thomas “TJ” McKenna is a clinical assistant professor in science education at BU Wheelock College of Education & Human Development. Dr. McKenna has spent the past 20 years as a scientist, educator (formal and informal), on-air science communicator, and faculty member focused on uncovering what works well in STEM education and how we can improve in scaling research and practice-based efforts. He is also the director of the Center for STEM Professional Learning at Scale. Dr. McKenna works at the nexus of policy, practice, and research in ways that deepen the understanding of how to make progress on the ecological problem of supporting teachers (preservice & in-service) in their efforts to offer high-quality science learning to all our students. He is also the founder and creator of Phenomena for NGSS, an educational website designed to support teachers in learning more about phenomena-based instruction.

PI:  Hao Yu 2024 AI and Education Initiative Doctoral Fellow
Co-PI: Margrit Betke, Co-Director, Artificial Intelligence Research (AIR)
Department Director of the MS in AI Program
Professor, Computer Science, CAS

Project Summary: Hao Yu and Margrit Betke are working on computer vision tools to see how student engagement predicts student outcomes in an intelligent tutoring system.

PI Bio: Hao Yu is a PhD Candidate in the Department of Computer Science at Boston University, advised by Professor Margrit Betke. Before joining BU, he obtained his B.Eng. in Computer Science at Zhejiang University. His research focus includes computer vision, machine learning, and human-computer interaction, as well as their applications in affective computing and educational technology, including student behavior analysis and content personalization for intelligent tutors.

PI:  Leslie Dietiker ,Ph.D., Associate Professor of Mathematics Education, Teaching and Learning, Wheelock College of Education & Human Development

Project Summary: LLMs can make lesson plans, but they don’t always follow the evidence based practices that we know make for really high quality instruction. Leslie Dietiker is exploring how LLMs can be used to make high-quality lesson plans.

PI Bio: Leslie Dietiker studies ways to improve the mathematical experiences of secondary students. As a former high school teacher, she focuses on teachers’ use of curriculum materials and their curricular decision making. Dietiker is the lead author on seven mathematics textbooks for middle and high school through CPM Educational Program and is an elected executive member of the International Society for the Design and Development of Education (ISDDE). Dietiker is interested in the quality of mathematics curriculum materials, curricular decision making, mathematics teacher design work, and mathematical stories.

PI:  Kathy Kim , Ph.D.,Program Director, Language Education
Clinical Assistant Professor, Wheelock
Co-PI: Zachary Rossetti , Ph. D.,Associate Professor, Special Education, Wheelock
Co-PI: Manuel Ramirez , Ph.D. student in Educational Studies with a specialization in special education, Wheelock

Project Summary: We have a few different projects looking at automatic translation technology as it relates to multilingual students. Kathy Kim is building tools for evaluating language proficiency in students who are learning English. Zach Rossetti and Manny Ramirez are looking at how translation tools can empower families to advocate for students’ needs.

PI Bio: Dr. Kathy MinHye Kim is a clinical assistant professor and program director of language education at BU Wheelock College of Education & Human Development. Her work specializes in understanding how adult learners encode and retain second language (L2) knowledge. In particular, she examines how L2 learners develop implicit and explicit knowledge and factors that mediate their development; these factors include situational features (e.g., input modality), sleep conditions, and individual differences in cognitive (e.g., intelligence, working memory capacity) and biographical (e.g., education level and reading experiences) variables.
Dr. Kim is highly engaged in longitudinal research; in her recent work, she examined the longitudinal associations between linguistic knowledge types (e.g., explicit and implicit L2 knowledge) and how language engagement in a study aboard context impacts the knowledge interface. Currently, she is also working on exploring the development and evaluation of web-based linguistic measures and examining the construct validities of these measures.
Dr. Kim has received grants from the National Science Foundation (NSF), the International Research Foundation for English Language Education (TIRF), and the National Federation of Modern Language Teachers Associations/Modern Language Journal (NFMLTA/MLJ). She holds a PhD in second language studies from Michigan State University with an interdepartmental graduate specialization in cognitive science.

PI:  Erin Barno 2024 AI and Education Initiative Doctoral Fellow
Co-PI: Greg Benoit Assistant Director of the Earl Center for Learning and Innovation
Lecturer, Mathematics Education, Wheelock College of Education & Human Development

Project Summary: One of the big pain points in our college and we suspect many other colleges is that it is time consuming and expensive to get student teachers hands on experience. There is never enough time, and it is hard to orchestrate the learning experiences we might want. We see AI simulated teaching experience as a really rich opportunity for research and development, an area where BU could be well positioned to lead.

PI Bio: Erin Barno, Ph.D. candidate, bridges her mathematics education background with educational technology design, development, and research. Erin’s research informs how educational technology can create opportunities to learn how teachers develop an awareness of their choices and changing critical consciousness. Due to the persistent message that mathematics teaching and learning is politically neutral, her work is pivotal towards analyzing mathematics teacher learning to make sense of if and how equitable mathematics teaching is taken up before and while in the classroom. As part of this project and the MIT Teaching Systems Lab, Erin is currently exploring applications by which artificial intelligence can be coupled with the knowledge and collaboration brought by teacher experts to create generated feedback for novice teachers engaging in digital clinical simulations.

PI:  Michael Chang ,Ph.D.,Assistant Professor at Boston University Wheelock College of Education & Human Development, Faculty Fellow in the Faculty of Computing & Data Sciences (CDS), and Assistant Director at the Earl Center for Learning & Innovation

Project Summary: Michael Chang developed a course on co-design in AI and Education, exploring teachers and learners can participate in building the future of AI and Education, rather than having that future happen to them.

PI Bio: Dr. Michael Alan Chang is an assistant professor at Boston University Wheelock College of Education & Human Development, a faculty fellow in the Faculty of Computing & Data Sciences (CDS), and an assistant director at the Earl Center for Learning & Innovation. Dr. Chang is a learning scientist and computer scientist who envisions AI-supported possibilities for teaching and learning that go outside the dominant, status quo instructional practices of schooling. He builds on ethical, relational, and speculative approaches to participatory design and closely partners with students, their families, and their teachers.
Before coming to BU Wheelock, Dr. Chang was a postdoctoral researcher at the University of California Berkeley where he worked with two major projects: the NSF Institute for Student-AI Teaming (iSAT) and the Center for Integrated Research on Computing and Learning Sciences (CIRCLS). He led the development of the Learning Futures Workshops, a 3-years co-design effort that worked with youth, teachers, and families. His design work has directly led to the design and implementation of novel AI-based collaborative learning tools that support equitable and democratic outcomes of schooling. This tool, the Community Builder (CoBi), has been deployed in public middle and high schools across the country.
As a doctoral student, Dr. Chang designed and implemented privacy-preserving computing systems, studied system-level support for large scale training of deep neural networks, and developed approaches to automatically deploy and maintain “microservice” computing clusters. He publishes to a variety of different fields, spanning the learning sciences, AI for Education, human-computing interaction (HCI), artificial intelligence, and distributed systems and privacy preserving computing. His work has appeared in many publications, including Conference on Human Factors in Computing Systems (ACM CHI), International Society of the Learning Sciences (ISLS), Artificial Intelligence in Education (AIEd), and Operating Systems Design and Implementation (OSDI).