AI-Driven Tools for Teamwork and Reflective Practices
Our team has developed an AI-driven educational tool to enhance STEM students’ teamwork and reflective practices, deployed to nearly 700 students at Cornell University in Fall 2024 and Spring 2025. This tool, funded by Cornell’s Center for Teaching Innovation Innovation Grant, provides personalized feedback and structured reflection prompts to improve collaboration skills.
This research explores both survey design and how students engage with AI-based educational tools. From an survey perspective, key questions include how AI-generated reflective prompts can be tailored to maximize engagement and high-quality data collection, as well as how AI-driven tools can bridge the gap between traditional surveys and interviews in education research.
We also examine how students interact with AI-based educational tools through the lens of activity theory. A core question is how students conceptualize and respond to AI-generated feedback in the context of their collaborative learning experiences. This research considers how AI-mediated reflections shape students’ understanding of teamwork, decision-making, and group dynamics. By studying these interactions, we aim to better understand how AI can serve as a meaningful mediator in reflective learning and teamwork development.
Investigating Faculty Pedagogical Practices in Teamwork
Our team developed the Instructor Mindsets on Pedagogy and Attitudes towards Collaborative Teaming in Engineering (IMPACT) survey. This tool is designed to capture a wide range of engineering instructors’ perspectives across the United States, shedding light on the challenges and motivations surrounding the integration of teamwork into their curricula.
The respondents of this survey are engineering educators who have instructed an engineering course incorporating teamwork in any form within the past three years. This encompasses professors and instructors of all ranks across various disciplines and types of institutions (public, private, 2-year, 4-year, etc.), providing a comprehensive view of engineering instructor motivations and experiences across the country.
In the survey, participants are initially asked to reflect on a specific course that incorporates teamwork in any form. Keeping this in mind, instructors respond to a series of survey items grouped into six primary themes: (1) internal mindsets of student learning and intelligence, (2) external institutional and course contexts shaping the teaching and learning environment, (3) social support and challenges for instructors, (4) current teaching practices and experiences in the classroom, (5) motivations and attitudes towards teamwork, and (6) demographic information.
We conducted the survey with over 100 faculty members nationwide, representing a variety of institution types. We found that over 75% of the instructors reported that students feel safe bringing team-related issues to the instructional team; however, nearly the same percentage (72%) felt ill-equipped to effectively resolve these conflicts. This discrepancy underscores the pressing need to develop a more comprehensive and use-inspired understanding of teamwork, enabling instructors to facilitate it more effectively and equitably in their classrooms. Similarly, instructors reported that their students require assistance when it comes to teamwork skills: 62% agreed that students need guidance on how to work effectively with others. And 44% found it too difficult to assess individual contributions when students work in teams.
Course Redesign for Biomedical Engineering Design Education
Through the Cornell Course Redesign Initiative to Support Teaching for Engaged Learning (CRISTEL), we are transforming biomedical engineering design education by embedding reflective practices that cultivate expert-like design thinking. This effort includes redesigning a four-week design project in the introductory BME course and overhauling the BME 2080/2081 sequence.
Global Collaborations in Engineering Education
We are developing international interventions to improve student resilience and mental health in group-based assessments through the Cornell-QMUL Global Strategic Collaboration grant. Additionally, past work through the Cornell-UCL Global Strategic Collaboration grant has explored cross-institutional partnerships between the UK and US in engineering education.
Using AI-Generated Synthetic Data in Educational Research
Our research examines how synthetic data can expand qualitative research opportunities. This includes developing an observational tool to analyze how students respond to real-time feedback in design studios and using AI-generated “synthetic” data to train NLP models for qualitative analysis. Our findings demonstrate the potential of synthetic data in educational research, particularly when real data for NLP training is limited.
To learn more about our research projects and or to join our efforts, please reach out to Alexandra Werth for inquiries or further information.