coLearn-AI: Collaborative Learning with AI Support

coLearn-AI is a web-based learning platform designed to support small-group, collaborative learning with structured activities, guided assessment, and AI-assisted feedback.

It is inspired by evidence-based pedagogies such as POGIL, team-based learning, and case-based instruction, and is designed to work equally well for:

  • in-class group activities
  • formative quizzes
  • summative tests
  • AI-guided reflection and follow-up

The system emphasizes thinking together, not just submitting answers.

Some Useful LInk

Core Goals

  • Support collaborative problem solving in small groups (1–4 students)
  • Make activities explicitly structured, not ad-hoc
  • Allow instructors to author once and reuse across contexts
  • Provide immediate feedback without replacing instructor judgment
  • Enable AI-assisted guidance while preserving academic integrity

At a high level:

  1. Instructor authors an activity
    • Using a simple, readable text-based format
    • Stored in Google Docs or Sheets
  2. Instructor launches the activity
    • Students are grouped automatically
    • Each group gets its own activity instance
  3. Students work collaboratively
    • One active student at a time (rotates automatically)
    • Others observe and discuss
  4. Responses are saved incrementally
    • Completed sections lock
    • Progress is preserved
  5. AI support (optional)
    • Evaluates completeness
    • Generates follow-up prompts
    • Never overwrites student work

Key Design Choice: One Activity = One Group

Unlike many LMS tools:

  • Each group runs its own instance of the activity
  • Responses are group-level, not individual
  • The system tracks:
    • progress
    • completion
    • participation
    • active student rotation

This models how collaborative learning actually happens in the classroom.

Authoring Format (Instructor-Friendly)

Activities are written in a structured markup format that is:

  • readable as plain text
  • easy to version and revise
  • parseable by the system
  • independent of the UI

Example

\title{Clinical Reasoning: Chest Pain}
\studentlevel{Medical School – Pre-Clinical}
\activitycontext{
In this activity, your group will reason through a brief
clinical scenario involving chest pain.
}

\questiongroup{Initial Assessment}

\question{
A 54-year-old patient presents with chest pain.
List three immediate questions you would ask.
}
\textresponse{5}

\sampleresponses{
Onset, character, radiation of pain;
associated symptoms such as dyspnea or diaphoresis;
past cardiac history.
}

\feedbackprompt{
Does the explanation justify urgency and diagnostic value?
}Code language: JavaScript (javascript)

Why this matters

  • Faculty can focus on pedagogy, not UI mechanics
  • Activities can be:
    • collaborative
    • quiz-like
    • test-like
  • The same activity can run:
    • in class
    • asynchronously
    • with or without AI assistance

Two Modes: Activities and Tests

Collaborative Activities

  • Students work in groups
  • One active participant at a time
  • Discussion encouraged
  • AI can prompt deeper reasoning

Tests

  • Individual
  • Timed or untimed
  • No AI assistance
  • Same authoring format

This allows instructional continuity across learning modes.

Why This Matters

  • Encourages clinical reasoning, not memorization
  • Makes thinking visible
  • Supports formative assessment
  • Reduces grading load without automating judgment
  • Preserves instructor control
Scroll to Top