MIT STEP Lab Presents

Collaborative AI for Learning (CAIL)

Overview

Collaborative Artificial Intelligence for Learning (CAIL) is a research and design project focused on integrating innovative AI tools in the classroom to support learning in collaborative groups. 

 

Students interact with AI-powered conversational agents in group work and discussions. The agent is envisioned as a peer who would promote deeper thinking on the topics instead of an efficiency tool. We also aim to provide real-time information to teachers and students to support teachers’ formative assessment and students’ self reflection.

As part of the project, we are building out CAILA (Collaborative Artificial Intelligence for Learning and Analysis) to process the real-time data while keeping with human- and child-centered design principles and keeping humans in the loop.

 

To date, we have created conversational agents to work with high school students in PBL data science workshops and investigated potential roles the agents could play in team-building exercises. 

Our lines of inquiry span from the personas and roles of the agents to analytic tool development to formative and reflective assessments for learning. 

If you are interested in learning more about our project, please reach out to our lab and mention “CAIL” in your email!

teachers

Research

Our team will be presenting at the poster session and running a workshop at Psychonomic Society. Please see a copy of the poster here: 

Lin, G. C., Hanks, B., Anderson, E., Fenech, M., & Farid, A. (2024, November). Collaborative AI for learning and analysis: Establishing an approach to leverage LLMs for analysis of data generated in learning contexts [Poster Session]. Psychonomic Society, New York, NY.

Psychonomic-CAILyze finalized

Project Specifics

Audience

Content Area

Project Contact

eanderson@education.mit.edu