Pair-Up
Teacher Interface for AI-Powered
Student Pair Suggestions
I contributed to a research project in Summer 2022 as an undergraduate research intern at Carnegie Mellon’s Human-Computer Interaction Institute in Pittsburgh, Pennsylvania.
This was my first real-world UX project as a design student and my first time recruiting/interviewing people to evaluate a (pre-ChatGPT) AI prototype web interface designed to help teachers pair up students in the classroom to tutor each other.
Role
UX Researcher
Team
2 Undergrad Interns
4 Researchers
Tools
Figma, Miro, Google Docs
Duration
10 Weeks
Process
Recruiting, Usability Testing, Affinity Mapping, Concepting
SCENARIO
Teacher Overload in Busy Classrooms
You’re teaching a math lesson in the classroom and some students are struggling to solve the problems.
So you pair up all the students to tutor each other.
But with so many students to monitor, it’s difficult to know if every pair is effective and staying on task.
DESIGN GOAL
How might we ensure the interface helps teachers
create effective student pairs for struggling students?
THE EXISTING PROTOTYPE
AI-Powered Student Pair Suggestions for Teachers
In Pair-Up’s user interface, each student is represented by a square.
With a simple click, teachers can choose a student (the solver) and see AI suggestions for pairing them with another student (the tutor).
There were two main student pairing options:
🎲 Pair Randomly: Pairs students at random.
☯️ Pair by Different Knowledge: Suggests student tutors who can help another student (the solver).
Pair Up’s AI was informed by data from a separate student interface where students solved math problems.