Tracy Kidder's Mountains Beyond Mountains is a favorite in English, social studies, and public health courses because it pairs a gripping biography with hard questions about poverty, medicine, and responsibility. That richness also makes the essays it produces slow to grade, since students may write about Paul Farmer's character, the Haitian proverb behind the title, or the ethics of treating drug-resistant tuberculosis in poor communities. A teacher with 120 papers can spend entire weekends writing comments that say much the same thing. AI grading tools built around a rubric give educators a faster first pass without giving up their own judgment.

A typical response to this book falls into one of three patterns: a plot summary that admires Farmer, a thesis about global health inequity supported by two or three scenes, or a reflective piece that drifts into the student's own opinions on charity. Each pattern needs different feedback, and a rubric that separates argument, evidence, and analysis helps the grader see which one a student has actually written. When an AI tool applies that rubric consistently, the teacher can spend attention on the papers that need real conversation instead of the ones that simply need a score.

Consistency is the quiet benefit of this approach. By paper eighty, most graders hold the early essays to a slightly different standard than the late ones, and a student who wrote about Farmer's clinic in Cange, Haiti on Tuesday night may be judged more harshly than one graded on Saturday morning. An AI tool that reads every essay against the same descriptors does not tire and flags the same weakness the same way each time. Teachers can then adjust scores where context matters, such as for a student who wrestled honestly with a difficult passage.

What AI Can and Cannot Judge in a Book-Based Essay

AI is strong at tasks with clear criteria, such as whether a thesis takes a position, whether quotations are introduced and explained, and whether each paragraph stays on a single point. It is less reliable at deciding whether a student's reading of Farmer's choices is original or unfair to the book's nuance, which is where a teacher's knowledge of the class discussion matters most. For that reason, most schools treat AI output as a draft evaluation that a teacher reviews and adjusts before students see it. This keeps the efficiency while leaving final authority with the person who knows the students.

• Name the specific claim the student makes about Farmer, Kidder, or global health instead of praising the essay in general terms.
• Point to one quotation or scene that supports the claim and identify one that is missing or underused.
• Explain in plain language how the analysis could move beyond a summary of what Farmer did.
• Identify a counterargument the essay ignores, such as critics who questioned the cost of Farmer's approach to treatment.
• Offer one revision step the student can complete in about twenty minutes.

Building the Rubric Before You Run Any Essays

The quality of AI feedback depends almost entirely on the rubric behind it, so the most valuable hour in the process is spent before any essay is uploaded. Write criteria that match your actual assignment, for example a thesis that addresses a question about structural inequality, evidence drawn from at least three sections of the book, and analysis that explains why a scene matters rather than retelling it. Describe what a strong, adequate, and weak response looks like for each criterion in concrete terms. Vague labels such as "good analysis" produce vague comments, no matter who or what applies them.

It helps to test the rubric on three or four sample essays you have already graded by hand. If the AI feedback matches your instincts on a strong paper, a middling paper, and a weak paper, the criteria are probably specific enough. If it overrates a summary that merely sounds confident, tighten the language around analysis and add a line that explicitly separates description from interpretation. A few rounds of this calibration usually pays back many times over across a full class set.

Keeping the Teacher in the Loop

Mountains Beyond Mountains invites strong feelings, and students sometimes write passionately about people and places in ways that deserve a human response. A teacher reading AI-generated feedback can spot the essay where a student's real engagement with Haiti or Rwanda is buried under clumsy structure, and can add a personal note that encourages the student to keep going. That human layer is what turns efficient grading into good teaching. Automated comments handle the repetitive mechanics, which frees the teacher to respond to ideas.

Many educators also use the AI evaluation as a starting point for class-wide instruction. If twenty essays share the same weakness, such as quoting Kidder without explaining the quotation, that pattern becomes the focus of the next mini-lesson. The grading process then informs teaching rather than ending with a gradebook entry. Over a unit, students see their feedback connect directly to what happens in class, which makes revision feel more purposeful.

Setting Realistic Expectations for Turnaround Time

Faster grading does not mean careless grading, and teachers should decide in advance what the saved time will be used for. Some use it to return essays within two days instead of two weeks, since feedback delivered while the book is still fresh in students' minds is far more likely to be read. Others reinvest the hours in conferencing with students who are struggling or in giving longer comments on the best essays. Either choice is defensible, but it works better when it is planned rather than discovered later.

Schools that adopt this approach for a reading unit typically start small, with one class set and one clear rubric, before expanding to a team or department. That limited start lets teachers see where the AI feedback needs adjusting and build trust in the process. By the next time the book comes around, the rubric is refined, the comment style is familiar, and the grading workload for the unit has shrunk noticeably. The essays themselves, and what students learn from writing them, are the real beneficiaries.