AI Writing Feedback Generator for Math Class

The AI can read and evaluate written explanations that include mathematical notation pasted as text. It assesses the quality of the surrounding explanation and reasoning — whether the student explains why each step follows from the previous one, whether notation is used correctly, and whether the conclusion follows from the work. Typed notation (LaTeX or standard text) can be pasted; the teacher checks the math.

Yes. Show-your-work responses are well-suited to rubric assessment. A rubric for a multi-step problem might assess: Identifies the correct method, Sets up the equation correctly, Executes steps without arithmetic error, Arrives at correct answer, and Explains the result in context. The AI drafts feedback on each criterion and points to where the student's reasoning breaks down — not just whether the final answer is right. The teacher verifies the math and sets the marks.

Yes. Many standardised tests (state exams, AP) use rubric-scored constructed response questions. The draft gives a structure for rubric-level comments on practice test responses.

The tool penalises nothing, because it assigns no scores. The draft feedback treats each criterion independently: if notation is a separate rubric criterion, notation issues are noted there, while valid problem-solving steps are acknowledged under the reasoning criterion. A student who gets the right answer through valid reasoning but presents it sloppily can get a comment that says exactly that, and the teacher, reading the work, decides what each criterion deserves.

Yes. Statistical analysis reports — common in AP Statistics and college introductory stats — are a strong use case. Your rubric can include criteria for correct hypothesis identification, appropriate test selection, correct calculation procedure, p-value interpretation, and conclusion in context. Check for the common error of students concluding “the null hypothesis is true” rather than “we fail to reject the null hypothesis.”