Generative AI in Business Education: From Academic Integrity Panic to Judgement-Based Assessment

Generative artificial intelligence has heightened concerns about academic integrity in business education, but its deeper challenge lies in the validity, equity, and professional judgment of assessment. AI tools can now produce plausible case analyses, SWOT/PESTLE reports, market summaries, business plans and strategic recommendations. This raises a fundamental question: what are business schools assessing when polished managerial analysis can be machine-assisted or machine-generated? This analytical note argues that business education should move beyond AI detection and academic integrity panic towards judgement-based assessment. Drawing on HEPI/Kortext and Jisc evidence, it further argues that AI policy must recognize uneven levels of student use: avoidance driven by fear or unclear guidance; cautious experimentation; over-reliance that weakens learning; and strategic use by already confident students. Reflective AI disclosure, process evidence, oral defense, critical AI literacy and stakeholder-based decision tasks can help preserve trust while preparing students for AI-mediated professional practice.

Generative AI has changed the conditions of business and management education. Students can now use AI tools to draft case-study analyses, generate SWOT and PESTLE frameworks, prepare market summaries, produce business plans, develop entrepreneurship pitches and simulate strategic recommendations. This is especially important in business education because graduates will enter workplaces where AI-supported analysis is increasingly the norm, but where responsibility for judgment, risk, and ethical consequences remains human. The immediate institutional concern is academic integrity: whether students are submitting work that is partly or wholly AI-generated. That concern is legitimate. Yet a narrow focus on cheating risks missing the more important pedagogical issue. Generative AI exposes a weakness in business school assessment: many tasks rely heavily on polished written outputs rather than visible managerial judgement, evidence use, ethical reasoning and decision accountability.

Recent evidence shows that AI is no longer marginal in higher education. The HEPI/Kortext Student Generative AI Survey 2025 reported that 92% of UK students used AI in some form and 88% used generative AI for assessments. Freeman, [1] the report’s author, described behavioral change on this scale as “almost unheard of” and argued that generative AI is now deeply embedded in higher education. Importantly, the report does not justify a simple cheating-panic narrative. Janice Kay’s foreword notes varied AI use in learning and assessment, and limited evidence that students are simply using AI to “cheat and play the system”. This matters for business education because many business assignments are precisely the kinds of outputs that AI can now produce convincingly: structured reports, strategic options, market analyses and executive-style recommendations.

The most useful HEPI recommendation for business schools is that institutions should urgently “stress-test” assessments to check whether they can be completed easily using AI. For business education, stress-testing should not mean merely asking whether AI can generate an answer. It should ask whether the assessment still captures the intended learning outcomes: managerial judgement, evidence evaluation, ethical reasoning, contextual interpretation, and accountability for decisions. A generic market-entry report may be easy for AI to imitate. A task requiring students to justify assumptions, interrogate data quality, compare stakeholder consequences, defend trade-offs and explain how AI outputs were accepted or rejected is much harder to outsource without learning.

Jisc’s 2025 student perceptions work adds an important equity dimension. AI inequality is not only about access to tools. It is also about clarity, confidence, risk perception and the ability to use AI without weakening one’s learning. Jisc reports that some students avoid AI altogether because they fear that any use may be classed as cheating, lack clear guidance, do not know enough about AI, or feel they lack the skills to use it effectively for study. Others who initially relied heavily on AI for academic tasks reported diminishing returns, including lower-quality work and lower grades, prompting them to rethink their approach. These findings complicate the assumption that students are simply AI users or non-users. Students differ in their level of AI engagement, and each level carries distinct equity risks.

Business schools should therefore avoid a one-size-fits-all AI policy. Non-users may need reassurance, basic training and clear examples of legitimate use. Cautious users need discipline-specific boundaries. Heavy users need support to avoid dependency and to rebuild independent analytical habits. Strategic users need to be challenged to use AI critically rather than merely efficiently. Without this differentiated approach, AI may widen inequalities between confident students who know how to use tools well and anxious students who avoid them, misuse them or rely on them in ways that damage learning. These varied patterns of avoidance, cautious use, over-reliance and strategic engagement show that AI inequality is not simply an access issue. Table 1 translates these patterns into policy implications for business and management education.

Table 1:Levels of student generative AI use and implications for business education policy.

Note: The table synthesises patterns reported in Jisc’s 2025 student perceptions work and translates them into implications for business and management education. The categories are analytical rather than fixed student types; students may move between them depending on task, discipline, confidence, guidance and assessment stakes.

The challenge is therefore not only whether AI use can be detected. Detection-led responses may have some role, but they cannot be the central pedagogical answer, especially when they generate mistrust or discourage legitimate formative use. They may also create anxiety for students who use AI legitimately for brainstorming, language support or formative clarification. A business school culture organized around catching misuse is unlikely to teach students the professional capability they will need: the ability to use AI critically, transparently and responsibly in conditions of uncertainty.

The more productive response is to redesign the assessment around judgement. Judgement-based assessment refers to tasks that require students to justify assumptions, evaluate evidence, defend trade-offs, consider stakeholders and remain accountable for decisions, including where AI has contributed to the process. In business education, the central competence is not merely producing a fluent report. It is making, justifying and defending situated decisions. A student who submits a polished marketentry strategy should be able to explain the evidence used, identify assumptions, evaluate alternatives, justify trade-offs, consider stakeholders and acknowledge ethical risks. If AI helped generate options or language, the student should be able to explain how the output was assessed, revised or rejected. This shifts the focus from product authenticity alone to decision accountability. Bearman et al. [2] argue that generative AI makes evaluative judgement increasingly important: students need to judge the quality of AI outputs, AI processes and their own work. Luo’s [3] review of GenAI policies in higher education assessment similarly shows that institutional framings of AI often revolve around originality and misconduct, while the deeper assessment question concerns what originality should mean in technologically mediated learning. For business education, originality should not be reduced to unaided text production. It should include responsible analysis, contextual judgement and defensible decision-making.

Several practical responses follow. First, business schools should use reflective AI disclosure. Students should not merely declare whether they used AI; they should explain how they used it, what they accepted, what they rejected, what they checked against evidence and how the final judgement remained their responsibility. Disclosure should support learning rather than operate as confession. Secondly, assessments should include process evidence: drafts, decision logs, annotated AI outputs, short oral defenses, group debriefs or viva-style questioning. These practices do not eliminate misconduct, but they reduce dependence on a final written product as the only evidence of learning. Thirdly, tasks should be more judgement-rich. At the module level, this means redesigning tasks so that the final report is only one part of the evidence, alongside process notes, justification of assumptions, source evaluation and a short defense of decisions. Students might compare AI-generated and human-generated recommendations, critique an AI-produced market-entry plan, identify hallucinated or weak evidence, or defend a final decision before a panel.

Business education should also teach critical AI literacy as part of professional formation. Students should learn that AIgenerated business analysis may be fluent but flawed. It may hallucinate sources, reproduce market assumptions, ignore local context, flatten stakeholder complexity or present generic recommendations with unwarranted confidence. HEPI’s call for balanced policy is useful here: students need support to develop AI skills while being educated about risks such as hallucinations, privacy and environmental concerns. UNESCO’s guidance similarly emphasizes human-centered governance and teacher capability [4,5]. Future managers need to know how to challenge AI outputs, not simply produce them faster. Lecturers also need institutional support. Staff cannot be expected to stress-test assessments, teach AI literacy, detect misuse, reassure anxious students and redesign tasks without time, guidance and workload recognition.

Generative AI should push business schools beyond the panic over academic integrity. If business education is understood as fluent report production, AI will primarily be seen as a threat. If it is understood as the development of managerial judgement, ethical reasoning and accountable decision-making, AI becomes a prompt to redesign assessment around what matters most. The task is not to make students afraid of AI, nor to encourage uncritical adoption. It is to help them become professionals who can use AI, question AI and remain answerable for decisions made with its assistance.

1. Freeman J (2025) Student Generative AI Survey 2025. Higher Education Policy Institute / Kortext. Jisc. (2025) Student perceptions of AI 2025.
2. Bearman M, Tai J, Dawson P, Boud D, Ajjawi R (2024) Developing evaluative judgement for a time of generative artificial intelligence. Assessment & Evaluation in Higher Education 49(6): 893-905.
3. Luo JJ (2024) A critical review of GenAI policies in higher education assessment: A call to reconsider the “originality” of students’ work. Assessment & Evaluation in Higher Education 49(5): 651-664.
4. Miao F, Holmes W (2023) Guidance for generative AI in education and research.
5. Miao F, Cukurova M (2024) AI competency framework for teachers.