The argument rests on a striking pattern. Across five major domains of classroom assessment — assessment literacy, self-assessment, peer and collaborative assessment, inclusive assessment, and AI-assisted assessment — research overwhelmingly asks what students perceive as fair, not how they reason about fairness or how they might be supported to develop that reasoning. Even literatures organized around ethics and critical consciousness, such as critical assessment literacy, locate the fairness capacity in teachers and experts. Developmental programs now exist for preservice teachers, notably work using the Classroom Assessment Fairness Inventory to help trainee teachers wrestle with tensions between differentiation and equity, or between standardization and equal treatment, but no parallel program exists for students themselves.

To make the case, Rasooli draws on theories of moral agency, which concern personhood, responsibility, and the capacity to reason and act as a moral being. Two conceptions of the self underpin the analysis. The humanist self presents a unified identity that makes moral choices and is held accountable for them, compatible with a socio-cognitive view in which moral agency operates through monitoring one’s conduct, judging it against personal standards, and generating affective self-reactions. The poststructuralist self, by contrast, is produced through discursive categories and networks of relations that make certain ways of thinking possible. Crucially, the article argues these models are complementary: humans navigate a plurality of competing categories and engage in boundary work to unify them into coherent action. Either way, moral agency emerges as a skill that can be trained through practice.

The five assessment domains reveal the same recurring gap. In assessment literacy, teachers’ ethical capacity is modeled and studied, but its effect on students goes unexamined; students’ fairness perceptions are instead rated against teachers’ or experts’ judgements. In self-assessment, research has focused on accuracy, showing that higher-performing students tend to underestimate and weaker students overestimate their performance, with bias treated as a calibration problem rather than an opportunity for students to take responsibility for honest self-ratings. Rasooli points to an unrealized route here: self-assessment could support moral agency in character and citizenship education, with students reflecting on dimensions of democratic participation ranging from critical community engagement to basic personal education.

Peer and collaborative assessment tells a similar story of reporting without reasoning. Students consistently report unfairness when they cannot choose group members, when workload is uneven, or when free-riding goes unaddressed, and they object to identical grades when individual contributions differ. But the proposed fixes — individual marks and peer grading — reproduce, rather than question, the underlying logic. Read through a poststructuralist lens, categories such as equality, individualism, and meritocracy form a network of relations that defines fairness as equal distribution and individual attribution. An alternative network built on participation, representation, and relational recognition could reframe the debate. Meanwhile, peer assessment research shows friendship bias and psychological risk: students over-score friends and may withhold honest reports out of loyalty or fear of retaliation, suggesting that explicit classroom conversations about these risks could cultivate a culture of psychological safety and moral growth.

Inclusive assessment offers both the sharpest illustration of the problem and a model worth building on. Synthesis research shows that students with disabilities often experience assessment itself as a site of exclusion, and many avoid accommodations they are entitled to, citing stigma, embarrassment, or a perceived unfair advantage over classmates. The accommodation model, critics argue, legitimizes a binary between disabled and non-disabled students, making normalcy a category that assessment actively stabilizes. Yet one recent study pointed toward what a moral-agentic agenda could look like: 139 students with disabilities generated their own proposals for fairer assessment, including multiple formats, flexible procedures, choice over how and when they are assessed, and simpler adjustment systems. Rasooli suggests extending this by working with all students to examine how medical and social models of disability shape their fairness conceptions.

It is artificial intelligence that has made the argument newly urgent. A systematic review of AI-assisted assessment identified three dominant discourses — risk-oriented moral panic, rule-based compliance exemplified by AI declaration forms, and assessment design focused on generating evidence of individual learning — and in all three, students appear as compliant subjects rather than active participants. The evidence of student confusion is mounting. Some students judged AI-based grading fairer than human grading; others considered improving an AI-generated draft acceptable while peers deemed it misconduct. A survey of 1,217 respondents found a substantial majority regarded incorporating AI into academic work as outright cheating, while researchers have documented a phenomenon called AI guilt, a moral discomfort at using AI for tasks traditionally done alone. A survey of perceptions of AI-giarism found a full standard deviation of variability across items, indicating that students draw the line between legitimate and illegitimate use in materially different places.

Beneath the confusion lies a reasoning-conduct gap. A scoping review of 38 studies found that students who judged a practice unethical nonetheless sometimes engaged in it, citing unclear institutional guidance, workload pressure, and the normalizing effect of peers doing the same. Yet the same review found students themselves expressing interest in developing their moral agency. Some students reasoned that their work should come from them, locating fairness in whether work represents their own self rather than merely whether it can be attributed to them — a concern distinct from the technical question of validity. Others weighed AI’s usefulness against fear of dependence, or deferred entirely to institutional policy. AI-assisted assessment has been described as a wicked problem in which security, learning, fairness, authenticity, and workload cannot be optimized by any single solution, which is precisely why the article argues that compliance instruments must be complemented by an educative layer that develops students’ fairness reasoning.

The prescription is concrete. Rasooli proposes structured deliberation exercises in which students respond individually to fairness dilemmas spanning group work, cheating, exams, grading, and feedback, share and compare their responses with peers, and reflect with instructors on the principles driving their similarities and differences. The goal is not to hand students definitive answers but to help them examine their tacit assumptions, notice where their beliefs and actions diverge, and appreciate the plurality of discursive categories — equality, individualism, meritocracy, and alternatives such as participation and recognition — that shape what fairness means in any context. Instructors should expect strong emotions in this process and treat them as part of the journey. The article also cautions that recognizing student agency does not guarantee it will serve fairness: students can resist tasks, cheat, or reject feedback to protect their egos or relationships, which is all the more reason to give them targeted opportunities to develop their moral reasoning. The next decade of fairness research, the article concludes, should treat students as co-reasoners in assessment — capable of deciding which courses of action are right amid competing normative choices, rather than only reporting what feels unfair.

Subject of Research: Student moral agency in classroom assessment fairness across five assessment domains including AI-assisted assessment

Article Title: Beyond passive rights-holders: toward student moral agency in assessment fairness

Article References: Rasooli, A. (2026). Beyond passive rights-holders: toward student moral agency in assessment fairness. Frontiers in Education, 11, Article 2011570. https://doi.org/10.3389/feduc.2026.2011570

Image Credits: AI Generated

DOI: 10.3389/feduc.2026.2011570

Keywords: assessment fairness, moral agency, classroom assessment, student voice, self-assessment, peer assessment, inclusive assessment, academic integrity, generative AI, assessment literacy, student agency, educational ethics