Several colleges and universities have selected Mountains Beyond Mountains as a common reading for incoming first-year students, and writing programs often build an essay assignment around it. The appeal is clear, since the book raises questions about service, privilege, and responsibility that students can connect to their own lives. The grading challenge is equally clear, because a single program might have thirty sections, a dozen instructors, and thousands of essays to evaluate. Consistency becomes a program-level concern rather than an individual one.

Common reading essays are typically the first college-level writing a student submits, which makes the stakes unusual. Students are anxious, instructors are calibrating expectations, and the essay may be used for placement or early advising decisions. If one section grades generously and another strictly, students notice, and the program's credibility suffers. A shared rubric applied consistently is the best protection against that kind of drift.

Writing program directors often hold norming sessions where instructors score sample essays together and discuss disagreements. These sessions are valuable but time-consuming, and they only affect the instructors who attend. AI grading tools add another layer of consistency by applying the agreed criteria to every essay in the same way. Instructors can still override scores, but they begin from a common baseline rather than from their own habits.

Designing a Prompt That Produces Gradable Writing

The prompt determines how easy the essays are to evaluate. A prompt such as "Discuss Paul Farmer" invites loose summary, while one that asks students to evaluate whether Farmer's insistence on treating every patient is a model or an exception creates a clear argumentative task. Specific prompts also make AI feedback more useful, since the rubric can reference the question directly. Spending an extra hour refining the prompt can save dozens of hours in grading.

• Ask a question with at least two defensible answers so students must take a position.
• Require evidence from specified portions of the book, such as Haiti, Peru, and the Boston setting.
• Include a clear length and format expectation so essays are comparable across sections.
• Ask students to address one possible objection to their argument.
• State how the essay will be assessed and share the rubric at the same time as the prompt.

Balancing Program Consistency With Instructor Autonomy

Instructors sometimes worry that a shared rubric limits their professional judgment, and that concern deserves a serious answer. The most workable approach fixes the core criteria across sections while leaving room for instructor-specific comments, weighting adjustments, and emphasis on particular skills. AI-generated feedback can serve as the common layer, with instructors adding context for their own students. This keeps the program standard intact without turning teachers into scorers.

Adjunct and graduate instructor teams benefit especially from this arrangement. Newer instructors often lack confidence about what a B essay looks like on a first-year assignment, and a shared rubric with clear descriptors gives them a reliable reference. The AI feedback shows how the criteria can be applied in practice, which also serves as informal professional development. Over time, instructors internalize the standards and rely less on the tool.

Using Aggregate Data to Improve the Program

When hundreds of essays are scored against the same criteria, the data reveals patterns that individual graders cannot see. A program might discover that students consistently score well on thesis clarity but poorly on analysis, which suggests that instruction should emphasize how to explain evidence. Another program might find that essays about the Peru chapters are stronger than those about Russia, prompting a conversation about which readings students need more support with. Such insights are rarely available when grading is done in isolation.

Program directors can use this information in curriculum planning, professional development, and communication with other departments. It also supports accreditation reviews, where evidence of consistent assessment and use of data for improvement is increasingly expected. The point is not to replace instructor judgment but to give it a stronger foundation. Aggregate patterns inform decisions that no single instructor could make alone.

Getting Started Without Overhauling the Program

A realistic starting point is a pilot with a few volunteer instructors in a single semester. Each pilot instructor grades one set of common reading essays using the shared rubric and AI feedback, then compares the experience with their usual process. Collect notes on time spent, comment quality, and student reactions, and use them to refine the approach. Small pilots build trust faster than program-wide mandates.

If the pilot goes well, expand gradually and share results openly with instructors. Emphasize that the tools are meant to reduce repetitive work, not to remove teaching judgment. Programs that adopt this approach often find that instructors have more energy for conferences and revision workshops, which are the activities that most improve student writing. That shift is where the real value emerges.