Teachers are often skeptical when they hear that AI can grade essays about poetry, and the skepticism is reasonable. Poetry analysis depends on interpretation, nuance, and an ear for tone, all things that feel distinctly human. Essays about the women poets in Catherine Reilly's Scars Upon My Heart make a good test case, since the poems are emotionally layered and the strongest student responses notice small details of language and structure.

The honest answer is that AI performs well on the parts of grading that follow a clear pattern and less well on the parts that require a reader's judgment about originality. It can reliably check whether a thesis makes an argument, whether quoted evidence is explained, and whether the essay addresses the criteria in the rubric. It is less dependable at recognizing when a surprising interpretation is actually insightful, which is why the teacher stays in charge of final scores.

What makes the difference is the quality of the rubric the tool is given. A vague instruction such as "grade the analysis" produces vague feedback, while a rubric that specifies evidence, explanation of poetic technique, and accuracy of historical context gives the system something concrete to measure against. Teachers who invest twenty minutes in writing sharper criteria tend to see far more useful comments across the whole class.

Where AI Helps Most on Poetry Essays

The biggest gains come from the first pass over a large stack. An AI system can read every essay, compare it to the rubric, and flag papers that never quote a line, that summarize instead of interpret, or that confuse the poet with the speaker. That sorting alone lets a teacher spend limited time on the essays that need human attention, such as a student who is close to a breakthrough or one who has quietly misunderstood the assignment.

• Checking whether each paragraph connects a quotation to a claim
• Spotting essays that summarize the poem instead of analyzing it
• Drafting passage-specific comments tied to rubric language
• Flagging unsupported historical claims about the war or women's roles
• Producing consistent scoring across large sections of the same assignment

Where Teachers Still Need to Lead

Poems from this anthology often depend on irony, restraint, or a bitter shift in the final lines. A student who reads a sarcastic poem as sincere has made an error that an automated reader might miss if the essay is otherwise well structured. Teachers who know the poems well are better placed to catch those misreadings, so a quick skim of flagged or borderline essays remains an essential part of the workflow.

There is also the question of voice. Some of the most memorable student essays take a risk, arguing for an unusual connection between two poems or challenging the common idea that women's war poetry is purely mournful. Those papers can look messy on a rubric, and a teacher is the right person to decide whether the risk paid off and deserves extra credit.

Setting Up a Rubric That Works With AI

Write criteria in observable language, describing what a reader can point to in the essay. Instead of "shows deep understanding," try "explains at least two ways the poet's word choice shapes the speaker's attitude toward the war." Observable criteria give both human and AI graders the same target, and they also make it easier to explain scores to students who ask why they lost points.

It helps to test the rubric on three or four sample essays before running the full set. Compare the AI's comments to what you would have written, and adjust any criterion that produces confusing or off-target feedback. This small calibration step usually takes less than an hour and prevents the frustration of discovering a flawed rubric after a hundred essays have been scored.

Being Transparent With Students

Students respond better to AI-assisted feedback when they understand how it is used. Explain that the rubric was written by their teacher, that the feedback is reviewed before it reaches them, and that questions about a score can always be taken to a person. This transparency builds trust and reduces the sense that essays about deeply human poems are being judged by something that cannot appreciate them.

It also models good academic practice. When a teacher shows how a tool supports rather than replaces professional judgment, students learn that technology can handle routine work while people remain responsible for meaning. That lesson is surprisingly relevant to a unit about poets who wrote precisely because no official account captured what they were seeing and feeling.