Hannie Rayson's Two Brothers gives students a lot to write about, including political responsibility, family loyalty and the way public decisions land on private lives. That richness is also why the essays are slow to mark, since every student seems to read the play a little differently. A teacher with thirty scripts and a weekend can easily run out of energy by the tenth paper. AI-assisted grading offers a way to keep feedback consistent across the whole stack.

The first step is deciding what the essay is actually assessing before any tool touches it. For a text response on Two Brothers, that usually means a clear contention, accurate use of scenes and dialogue, and some awareness of how the playwright shapes audience sympathy. When those criteria are written down in plain language, both teachers and software can apply them the same way each time. Vague criteria such as "good analysis" produce vague scores from everyone.

Next comes the rubric itself, which should describe what each performance level looks like in a play-analysis essay. A top-band paragraph might connect a specific moment of stage conflict to the writer's argument about duty versus conscience. A middle-band paragraph might retell the scene correctly but stop before explaining why it matters. Writing these descriptors out in advance lets AI feedback point to the exact gap between a student's draft and the next level.

What AI Can and Cannot Do for Play Analysis

AI grading tools are strongest at the repeatable parts of marking, such as checking whether a claim is supported, whether evidence is explained, and whether a paragraph links back to the question. They are weaker at recognizing a truly original reading of the play that departs from the classroom consensus. For that reason, the best workflows treat AI as a first pass that produces structured comments, with the teacher reviewing anything unusual. This keeps speed and professional judgment working together.

• Upload the essay prompt and rubric together so feedback matches the real task
• Ask for comments tied to specific criteria rather than a single overall score
• Review borderline essays yourself before releasing marks to students
• Spot-check a few scripts against your own marking to confirm alignment
• Save strong comment patterns to reuse across future Two Brothers units

Setting Up a Marking Workflow That Saves Time

A practical workflow starts with a small sample of essays marked by hand, ideally one strong, one average and one struggling script. Those samples show where the rubric language is unclear and where AI feedback might drift from what you expect. Once the sample looks right, the rest of the class can be processed in a batch with far less second-guessing. Teachers often find the calibration hour pays back many times over.

It also helps to separate feedback for drafts from grading for final submissions. Draft feedback can be more conversational and focus on one or two priorities, such as sharpening the contention or explaining quotations. Final grading should be tighter and tied to the scoring criteria. Keeping the two modes distinct prevents students from feeling overwhelmed by a long list of fixes right before a deadline.

Keeping Feedback Specific to the Play

Generic comments like "add more evidence" do little for a student writing about Rayson's characters. Useful feedback names the moment in the play that could strengthen the argument, or points out that a quotation has been dropped in without any explanation of how it supports the point. When the rubric and prompt are supplied, AI tools can produce this more targeted language at scale. Teachers should still edit comments that miss the nuance of a particular scene.

Another common issue is feedback that praises plot summary as if it were analysis. Students writing on drama often narrate what happens on stage because it feels safe, and a good comment will gently redirect them toward interpretation. A line such as "you describe the confrontation accurately; now explain what it reveals about the character's priorities" gives a concrete next step. That small shift often raises essays by a full band.

Measuring Whether the Approach Works

To judge whether AI-assisted grading is helping, track a few simple things over a unit: how long marking takes, how consistent scores are across classes, and whether students actually act on the comments. If revision drafts show clearer contentions and better-explained evidence, the feedback is doing its job. If students ignore the comments, the problem may be length or tone rather than accuracy. Adjust accordingly and keep notes for next year.

Over time, schools can build a shared library of rubrics and exemplar responses for Two Brothers that every teacher in the department uses. This reduces the gap between how different markers treat the same essay. It also gives new teachers a clear picture of what strong work looks like. The result is fairer grading for students and a lighter load for staff.