Tilburg.ai | AI as a critic

AI as a critic on student's assignments

Students may use AI during the assignment, but they must also get AI feedback on their work. They hand in both the assignment and the AI critic feedback, with a reflection on which feedback they integrated (or rejected) and why. This encourages critical engagement with AI output.

Step 1

Step 1: Pick an assignment

Before creating your AI critic, decide which assignment students will work on.

Choose the assignment

Pick an assignment you already use in your course or make a fictional one. Ensure it has clear criteria so the AI critic can give targeted feedback.

Step 2

Step 2: Create the critic chatbot in Tilly

Create a chatbot that acts as a critical reviewer for your specific assignment. Students will use it to get feedback before submitting. Use the video as a walkthrough of the same setup in Tilly.

The video above shows how to create and configure a feedback-style chatbot in Tilly; follow it alongside the steps below.

Go to Tilly and create a chatbot

Log in and navigate to the create page. Create a new chatbot with a clear name (e.g. “Essay Critic” or “[Course] Assignment Reviewer”).

Or start from our critic template (pre-filled name, description, and instructions you can edit).

Use our template

Define the critic instructions

In the chatbot instructions, specify that it should act as a critical reviewer. Example prompt:

You are a critical reviewer for [assignment]. Provide constructive feedback. Base your feedback on the rubric and the assignment criteria.

Optional: Upload assignment materials

Upload your assignment rubric, criteria, or sample solutions to help the chatbot give more aligned feedback.

Step 3

Step 3: Plan submission and evaluation

Decide how students will submit their work and how you will evaluate both the assignment and their use of AI feedback.

Submission format

Clarify what students must hand in: (1) the final assignment, (2) the AI critic feedback (e.g. screenshot or exported conversation), and (3) a short reflection on which feedback they integrated or rejected and why. Consider a single document with an appendix, or separate files.

Evaluation criteria

Evaluate both the quality of the final assignment and the student’s engagement with AI feedback. Did they integrate good suggestions? Did they reject unhelpful feedback with sound reasoning? Include a rubric item for “use of AI feedback” to make expectations explicit.

Optional

End-of-course survey on AI feedback

Optional: if reflecting on how students learned to work with an AI critic fits your course goals, add a short closing survey.

Align with course learning goals

If you want students to meta-reflect on using AI as a reviewer, not only on the assignment topic, say so in your outcomes or instructions.

End-of-course survey

Near the end of the term, invite a brief survey about how students used the critic chatbot and what they took away for future work.

Example questions you can adapt:

How did you use the AI critic during this course’s assignments, and what worked or didn’t?
When you chose to accept or reject specific feedback, what criteria did you use, and what would you do differently next time?
How did handing in critic feedback alongside your work change how you revised or planned your writing (or other deliverables)?

How you might use the results

Look for patterns in how students engage with AI critique to tune prompts, rubrics, or in-class guidance next time.

Still got questions? Take a look at the FAQ page.