Include an explicit rubric in GenAI feedback prompts, even in Zero-shot conditions¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The article recommends supplying a detailed, task-relevant rubric when prompting GenAI for feedback. Unlike prior studies that withheld rubrics, "the present study included a detailed and relevant rubric for providing feedback, even in the Zero-shot condition", which the authors credit for Zero-shot feedback matching teacher feedback quality, since referencing rubric criteria "is a key characteristic of effective feedback". The discussion frames prompt design as "a fundamental pedagogical strategy" rather than a technical adjustment.

Design Implications¶

Context¶

Requirements¶

• a rubric aligned with the assessment criteria for the writing task

Constraints¶

• the recommendation is grounded in this study's objective, Toulmin-based quality assessment and GPT-4o; the authors note earlier studies with other models and subjective criteria favoured teacher feedback

Target Learners¶

• higher education instructors and researchers designing GenAI feedback workflows

Target Learning Goals¶

• generating high-quality, criterion-referenced feedback on argumentative writing

Related Strategies¶

• Provide LLM graders with the full human rubric, accepted solutions, and common-mistake penalties in the prompt (rubric-enhanced prompting)
• Reserve LLM grading for short, well-structured tasks with explicit rubrics or exemplar answers, newer models, and stable API pipelines

Examples¶

-

Key Sources¶

• Farrokhnia, M., Latifi, S., Papadopoulos, P. M., Hogenkamp, L., Gijlers, H., Khosravi, H., & Noroozi, O. (2026). Generative AI offers more, but students revise less: comparing the effects of teacher and AI feedback on student essay revisions. International Journal of Educational Technology in Higher Education. https://doi.org/10.1186/s41239-026-00579-9