Skip to main content

AI Grading in Colleague AI: Teacher Guide

Introduction​

Grading can take hours, especially when you want to give each student meaningful feedback. Colleague AI is designed to save you time while keeping you in control. Once you create an assessment and rubric, the system reviews student work, generates suggested scores, and explains its reasoning. You remain the decision-maker: you can edit any score or feedback before sharing it with students.

This guide walks you through each step:

  • How to create and prepare an assessment.
  • How to run AI grading and interpret results.
  • How to give feedback while keeping your professional judgment central.
  • Tips for making the most of AI support in your classroom.

Key Features (What Teachers Get)​

  • Flexibility in student work – Students can submit in many formats: typed responses, PDFs, scanned handwriting, or cloud documents. You can use Colleague AI for essays, lab reports, problem sets, or creative projects.
  • Feedback with explanations – The AI doesn’t just give a number. It points out strengths and growth areas for each rubric category, so students understand why they received a score.
  • Teacher authority – Nothing is final until you approve it. You design rubrics, set weighting, and review every score and comment. Think of the AI as a helper, not a replacement.
  • Rubrics aligned with standards – You can use ready-made frameworks (Common Core, NGSS, Bloom’s Taxonomy, Depth of Knowledge, or “Reaching Standard”), or create your own scales.
  • Resubmission support – If you allow it, students can revise and resubmit work. This turns grading into part of the learning process instead of just a final judgment.

Creating an Assessment​

  1. Start with the task
    On your Classroom page, choose New Assessment. Write a clear description of what you want students to do. If you have worksheets, readings, or other handouts, attach them here so students see everything in one place.

  2. Attach a rubric

    • If you already have a rubric (Word, PDF, etc.), upload it under “My Document” and mark it as “Rubric.”
    • If you don’t have one yet, click Create with Claire. Claire can build a rubric for you based on standards, cognitive levels, or custom performance levels.

  3. Add answer keys or exemplars (optional, but powerful)
    Upload sample answers or model work. These examples help the AI calibrate its scoring and make the feedback more accurate.

  4. Set grading preferences
    Decide if students should see AI feedback right away (great for practice tasks) or only after you finalize grades (better for exams). You can also set the due date, mark whether the task counts in the gradebook, and choose how many attempts are allowed.

  5. Publish the assessment
    Review your setup. When you’re satisfied, click Create and switch the status to Active. Students will now see the assignment and can begin submitting their work.

Running AI Grading​

  1. Collect submissions
    Students upload their work in any supported format. Each student has a separate tab for easier navigation.

  2. Check submission status
    In your dashboard, look for color cues:

    • Purple = student has submitted.
    • Gold = A grade has been returned to the student.

  3. Review the AI results

    • You’ll see a provisional score in the “Grade by Claire” field.
    • A detailed grading report appears in the right-hand panel, showing how the AI applied the rubric.

  4. Edit before releasing
    As the teacher, you should:

    • Adjust scores that feel too high or too low.
    • Re-write or add comments in your own voice.
    • Ensure language is positive, professional, and appropriate for your students.

  5. Return feedback
    Once satisfied, click Return Feedback. Students will receive the score and feedback you approved.

  6. Support growth through resubmission
    If enabled, students can revise and resubmit. The AI can grade the new version and track growth over time. You decide whether the latest score replaces the old one or is added alongside it.

Best Practices and Tips​

  • Be specific in rubrics
    Clear descriptors lead to more reliable scoring. For example, instead of “shows strong understanding,” write “explains concept using correct terminology and provides at least one example.”

  • Match rubric to the goal
    Use Bloom’s Taxonomy for higher-order thinking tasks, Depth of Knowledge for problem solving, or simpler scales for skill practice.

  • Provide exemplars
    Sample responses not only train the AI but also help students understand expectations.

  • Watch for bias
    AI can sometimes misinterpret creative or nontraditional answers. If you notice patterns—like certain student groups getting inconsistent results—step in and adjust.

  • Be transparent with students
    Share the rubric before they start. Explain how the AI supports grading and that you still review every score. This builds trust.

  • Encourage self-assessment
    Ask students to use the rubric on their own work before submitting. If you enable instant AI feedback, they can compare their self-assessment with the AI’s suggestions.

  • Use data to plan lessons
    The Student Growth Insights dashboard shows class-wide trends, common errors, and individual progress. Use this to identify what needs reteaching.

  • Protect student privacy
    Always follow your school’s policies on student data. Avoid uploading sensitive personal details beyond what’s needed for grading.

Frequently Asked Questions​

What if the AI gives the wrong score?
You can always edit. It’s good practice to double-check a few responses before sending feedback. If you notice the AI repeatedly misinterprets something, adjust your rubric or add sample answers.

Can students see AI feedback before I finalize grades?
Yes, it’s your choice. For formative work, letting them see feedback right away encourages revision. For summative tasks, you may want to hold feedback until you’ve confirmed scores.

Does AI replace teachers?
No. The system is built on a “human-in, human-out” model. The AI suggests, but the teacher decides. Your expertise, judgment, and relationship with students remain central.

Conclusion​

Colleague AI is here to lighten the grading load, not take your place. By combining fast, standards-aligned scoring with your professional oversight, you can:

  • Give students timely, meaningful feedback.
  • Focus on teaching instead of repetitive marking.
  • Support growth through feedback and resubmission.

When used thoughtfully, AI grading helps create a classroom where assessment is both efficient and student-centered.