TRAIL
Team Research and AI Integration Lab
Studying how humans and AI can work together as teammates — not just tools — to enhance innovation, creativity, and inclusion in collaborative problem-solving.
A research platform by the Language and Learning Analytics Lab at UC Irvine School of Education
Research Focus
Experimental Simulation of Team Processes
Scalable, replicable experiments on team cognition and decision-making. TRAIL enables RCT-style studies that vary task structure, group composition, and AI teammate behavior. These simulations produce rich datasets for studying innovation, bias, and socio-emotional regulation in teams.
Socio-Cognitive Dynamics & Discourse Analytics
Modeling the cognitive, emotional, and social processes of teamwork. We analyze team discourse using computational linguistics and learning analytics to identify markers of productive, inclusive, and adaptive collaboration. Findings inform real-time feedback systems that help humans and AI coordinate effectively.
AI Teammates for Inclusive Collaboration
Understanding how AI can foster equity, trust, and belonging in teamwork. Generative AI agents participate as teammates in controlled studies. By varying AI persona and communication style, we study how different behaviors influence participation balance, trust, and inclusion in mixed human–AI groups.
The TRAIL Platform
A comprehensive web application for managing AI-powered collaborative tasks and research studies.
- Experiment Design Create, manage, and analyze human-AI collaboration experiments with flexible team structures
- Configurable AI Teammates Design AI agents with customizable personalities, communication styles, and behavioral parameters
- Real-time Collaboration Chat-based interface for synchronous team problem-solving with humans and AI
- Analytics Dashboard Rich datasets and discourse analytics for studying team dynamics and outcomes
Studies on TRAIL
One question. Three studies.
Not “How good is the AI?” — but “What does an AI teammate do to the team around it?” Each study swaps one human seat for one AI teammate and watches what changes in the talk, the bonds, and the decisions.
Read the room, or lead the room?
Does an AI teammate's persona and timing change team dynamics? They aren't cosmetic: the same contribution lands differently depending on the AI's stance and its sense of timing.
The Mountain Rescue
33 teams faced a moral dilemma — rescue a stranded climber in a storm, or not? Half the teams included an AI teammate; half were all-human controls.
Temporal HAT — Going Longitudinal
17 teams followed across 5 sessions, half with a verbose AI persona and half with a terse one. Do teams rebuild belonging over weeks — or does withdrawal become the new normal?
An AI teammate is never neutral. It reshapes who talks, how connected people feel, and how the team decides — for better and worse at once. Study the team, not just the tool.
Interested in Using TRAIL?
We're looking for collaborators and research partners to explore human-AI teaming. Request a demo to see how TRAIL can support your research.
Request a DemoSelected Publications
- Minds and Machines Unite: Deciphering Social and Cognitive Dynamics in Collaborative Problem Solving with AI LAK'24
- The AI Collaborator: Bridging Human-AI Interaction in Educational and Professional Settings arXiv 2024
- Read the Room or Lead the Room: Understanding Socio-Cognitive Dynamics in Human-AI Teaming LAK'26
- AI Teammates and Inclusion Analytics: Revolutionizing Equity in STEM Collaboration Society for Text and Discourse 2025
- Human-AI Collaboration and Culture For Equity and Inclusion (Half-Day Workshop) AIED 2025
- Beyond Tools, Toward Teammates: Human-AI Teaming for Inclusive Learning Analytics LAK'26
- Catalyzing Equity in STEM Teams: Harnessing Generative AI for Inclusion and Diversity Policy Insights 2024
- Measuring Inclusion in Interaction: Inclusion Analytics for Human-AI Collaborative Learning arXiv 2026
- Advancing Knowledge Together: Integrating Large Language Model-Based Conversational AI in Small Group Collaborative Learning CHI 2024
- Coordinating Minds: Human-AI Teaming and Team Shared Cognition AERA 2026
- TRAIL: An Experimental Platform for Human-AI Collaboration Research ISLS 2026
- Personalities at Play: Probing Alignment in AI Teammates arXiv 2026
About TRAIL
TRAIL (Team Research and AI Integration Lab) is a research platform developed by the Language and Learning Analytics Lab at UC Irvine School of Education.
Our mission is to understand how humans and AI can collaborate effectively as true teammates — moving beyond AI as mere tools to AI as active participants in creative problem-solving.
Through rigorous experimental methods and advanced analytics, we investigate the dynamics of human–AI teams to inform the design of more inclusive, innovative, and effective collaborative systems.
Our Team
Seehee Park
Graduate Student
School of Education
Jaeyoon Choi
Graduate Student
School of Education
Spencer JaQuay
Graduate Student
School of Education
Pedro de Bastos
Graduate Student
School of Education
Get in Touch
Interested in our research or potential collaboration?
lalalab.ucied@gmail.comLanguage and Learning Analytics Lab
UC Irvine, School of Education