Curating truth or simulating thought? The ethics of AI-generated case studies in business education

The first time I used artificial intelligence (AI) to generate a case study outline in January 2023, I felt both impressed and unsettled, having just started experimenting with ChatGPT, unsure of what to expect. I typed a simple prompt for a marketing case study and within seconds, a complete outline appeared: coherent, structured and pedagogically sound. That prompted me to pause, it was not just astonishment and shock that I felt, there was a quiet unease that followed: whose voice was this?

I had spent years writing and refining case studies: drafting outlines, checking context and aligning to learning outcomes. But this time, the development had been outsourced, leaving me holding a usable draft generated in under a minute. The convenience was staggering, as was the potential implication. This moment marked a fundamental shift in my role as an educator, transforming me from solitary author to something else entirely: an editor, a protagonist in my own living case study and a curator of content. If AI can simulate thought, then what does it mean to teach thinking using AI-developed material?

Much of the existing literature on AI in education has focused on questions of technological integration, learning outcomes or institutional policy (Gentile et al., 2023), while fewer works have examined the shifting epistemic role of educators in AI-mediated authorship. This paper extends that conversation by offering two original conceptual tools developed through practice: the Curated Authorship Model, which maps the evolving human–AI writing partnership, and the 4W’s Ethical Framework, a set of reflective questions for evaluating what educators gain, lose, neglect and newly access when using AI in pedagogical design. By centering intentionality over automation and emphasizing the educator’s role as a narrative ethicist, these tools contribute to a growing discourse on how knowledge is curated, disclosed and taught in an era where machines simulate thought, but only humans can exercise judgment.

Case-based pedagogy has long been a cornerstone of business education. From its origins at Harvard Business School, shaped by pioneers like C. Roland Christensen and David Garvin, the case method has emphasized learning through dilemmas: real-world problems framed for analysis, decision and discussion (Christensen et al., 1991). At its best, a case study simulates complexity, offering students the opportunity to apply theory, debate alternatives and wrestle with ambiguity. The case, in this sense, is not just a pedagogical tool but a narrative scaffold that shapes how knowledge is explored and evaluated (Garvin, 2003).

Traditionally, these cases were authored through a slow and deliberate process grounded in field research, interviews and synthesis. What Christensen called “discussion leadership” was anchored in careful editorial design and narrative sensibility (Christensen et al., 1991).

Generative AI tools like ChatGPT, Claude and Gemini now offer educators rapid ways to draft outlines, generate dilemmas and structure cases in seconds. Recent studies suggest that this technological acceleration is not only transforming workflows but also raising questions about authorship, academic integrity and instructional ethics (Ramos Salazar and Peeples, 2025; Dabis and Csáki, 2024). As Lang et al. (2024) argue, the craft of case writing is shifting from composing to prompting, from authoring to curating. This transformation demands new ways of thinking about labor, trust and the co-production of knowledge. The literature offers several complementary lenses to help frame the following emerging questions.

Michel Foucault’s “author function” suggests that the value of a text is shaped by how authorship is constructed and received, not simply by the content itself (Foucault, 1969). Roland Barthes’ famous “death of the author” pushes this further, questioning the primacy of singular voice in the creation of meaning (Barthes, 1977). Bruno Latour, through actor-network theory, reframes agency as distributed: authorship becomes an assemblage of tools, texts, contexts and collaborators, including, now, AI (Latour, 2005).

Erik Brynjolfsson, Kate Crawford and Cecilia Chan have written extensively on how automation reshapes creative labor. Their work alerts us to the risk of mistaking efficiency for insight and the danger of invisibilizing the cognitive and emotional labor still required by humans, even when AI generates the first draft (Brynjolfsson and McAfee, 2014; Crawford, 2021; Chan and Tsi, 2023).

Luciano Floridi’s model of curation as stewardship positions educators not merely as content transmitters but as guardians of the infosphere (Bawden and Robinson, 2017). Wayne Holmes builds on this, suggesting that AI in education demands a pedagogy of transparent mediation, where the how and why of knowledge production is just as critical as the content itself (Holmes, 2021).

Together, these thinkers offer a foundation for examining what happens when AI enters the creative and ethical terrain of education, not as a neutral tool, but as a co-actor in shaping how knowledge is produced, delivered and received (Table 1).

My first experience using AI in case study design was not strategic; it was personal. It began, quite literally, with a prompt and a sense of what if. I was not attempting to outsource my teaching; I was simply curious. I asked ChatGPT to generate a case for a marketing course, something about segmentation and audience targeting. The response came fast: a named protagonist, a plausible business dilemma and three thoughtful discussion questions. At first glance, any case writer would be able to critique and poke holes in the narrative and essence of this case, but the reality was that this ideation and example was created in a matter of moments with some simple text.

I was taken aback, not merely by the speed, but by the coherence of what I encountered. The structure was solid, the learning arc recognizable and the potential for classroom application evident. Yet, something essential was missing. The narrative lacked my voice. Its tone was flat, and its cadence failed to mirror my natural rhythm. Absent were the subtleties I intentionally weave into my work, the interplay of challenge and ambiguity that shapes my pedagogical approach. I have always regarded case writing as a deliberate, almost artisanal process: a form of sculpting in which tone, context and character coalesce to serve a precise learning intent. This experience, however, diverged sharply. It felt manufactured, mechanical rather than crafted.

As I revised the draft, I realized that something deeper was occurring. I was not writing in the traditional sense; I was curating. Traditional writing, as Emig (1977) argues, is a mode of learning, a deliberate act that externalizes thought. Yet I was not externalizing my own thinking. I was responding to a draft that had been externally generated. Rather than composing from scratch, I found myself engaged in a layered process of prompting, refining and redirecting. With each iteration, I added values, shaped tone and adjusted ethical context. I was not merely correcting errors; I was re-shaping meaning. AI remained a tool, but it was increasingly functioning as my copyeditor. In turn, I was authoring the dialogue, grounded in my experience and pedagogical intent. From start to finish, the content, intent and process lead into a working case, but how I got there was completely different.

The speed was exhilarating, and the scaffolding proved impressively functional. Yet, in the process, I lost something vital: my voice. I lost the slow, recursive thinking that typically defines my pedagogical approach. More significantly, I began to recognize a shift in my role. I was no longer the solitary author; instead, I had become an editor, an orchestrator, a curator of intellectual tone. I had assumed the position of an ethical filter at the end of a production pipeline, the final hand shaping what students would ultimately encounter.

That realization prompted the deeper questions that now frame this paper. What exactly am I gaining by integrating AI, and what might I be losing? Which pedagogical muscles am I neglecting to exercise? At the same time, what new possibilities, audiences, formats, and languages am I accessing through this tool? These ongoing tensions led me to develop two conceptual frameworks that have since guided my practice:

1. The Curated Authorship Model, which maps the evolving relational dynamics between educator and AI throughout the content creation process.

2. The 4W’s Ethical Framework, a reflective tool for identifying gain, loss, neglect and access when incorporating AI into pedagogical design.

Neither framework is intended to be definitive, but both are deeply intentional. That, I have come to understand, lies at the core of engaging with AI as an educator: the goal is not mere efficiency, but purposeful intentionality.

As I began experimenting with generative AI in my case writing, I noticed a fundamental shift in my creative process and I was no longer starting from a blank page. Instead, I began with AI-generated outlines, character prompts and scenario templates: structural designs I would then revise through my own pedagogical lens. This was not passive work, it was iterative, intentional and deeply reflective. This shift echoes recent findings that human–AI collaboration often unfolds in patterned, time-based cycles of prompting, editing and epistemic control (Yang et al., 2024).

My first step was to systemize what I was able to extract from the AI models. I did this by creating a prompt template that followed a systematic process to have the AI successfully generate a viable case study with one large prompt. To make sense of this evolving workflow, I developed what I now call the Curated Authorship Model, a conceptual tool that maps the relational dynamics between human educators and generative AI in the co-creation of educational content.

This model departs from static conceptions of authorship and instead aligns with evolving traditions in remix pedagogy, adaptive curation and reflective instructional design. While remix pedagogy emphasizes playful reconfiguration, the Curated Authorship Model foregrounds accountability and intentionality in the co-construction of meaning. McMahon and Firestone (2024) suggest that AI integration need not replace foundational teaching practices; rather, it can remix and extend them, revitalizing instructional strategies by layering in new modalities while maintaining pedagogical coherence. Similarly, adaptive curation, as explored by Sibley et al. (2024), positions the educator as an active orchestrator who filters, sequences and contextualizes AI-generated content to meet learner needs. In contrast to models of instructional design that prioritize clarity and alignment alone, Curated Authorship introduces an ethical reflexivity, characterized by a deep concern for how knowledge is framed, whose voices are centered and what forms of pedagogical labor are either revealed or obscured in the process.

At its core, Curated Authorship refers to the intentional, iterative shaping of AI-generated outputs by a human educator who assumes final responsibility for pedagogical coherence, ethical alignment, and contextual integrity. It represents an epistemic shift from author as originator to educator as epistemic filter.

The Curated Authorship Model is composed of four interdependent and overlapping phases:

1. AI narrative generation.

In this initial phase, the AI produces content ranging from case outlines to dilemmas and character scaffolds. The output is fast and generative but often lacks nuance, specificity or ethical depth. This stage provides structure, not authorship.

1. Human curation.

Here, the educator steps in not merely to revise, but to re-embed voice, cultural context and narrative intention. Curation includes aligning the draft with learning goals, refining tone and addressing representational choices. It is not editorial clean-up; it is epistemic authorship.

1. Collaborative iteration.

This is the dynamic, dialogic exchange between human and AI. Prompts evolve. Outputs are tested and refined. Each loop generates new narrative possibilities. I might ask the AI to “add tension,” “include a stakeholder from the Global South” or “reframe this from a DEI lens.” Creativity emerges not from the AI alone, but from the feedback system.

1. Ethical accountability.

Surrounding and informing every phase is a reflective interrogation: Should this be taught? What is the positionality of this case? Who is visible and who is erased? This phase centers the educator as a moral agent, a narrative ethicist responsible for what the AI cannot know, namely, the stakes of representation (Table 2).

To clarify how Curated Authorship differs from related frameworks, it is necessary to distinguish it conceptually from overlapping traditions and to specify the boundaries that define when curation becomes authorship (Table 3).

Curated Authorship becomes distinct from simple editing or facilitation when the following three threshold conditions are met:

1. Substantial revision – more than 30% of the AI-generated content is revised for narrative depth, ethical clarity or pedagogical alignment.

2. Ethical filtering – the educator actively assesses the representational implications of AI output, including bias, visibility and inclusion.

3. Pedagogical intentionality – final outputs are shaped in alignment with specific learning goals or instructional objectives.

When these criteria are met, the educator is no longer an editor or prompter, but an epistemic author, curating knowledge with responsibility and intentionality. Although there is no legally mandated “30% rule” in copyright law, this figure serves as a useful heuristic for substantial transformation. It reflects not only informal benchmarks for originality in creative work, but also academic standards for plagiarism, where similarity scores exceeding 15–25% often prompt scrutiny. This threshold has gained new relevance as AI systems challenge the boundaries of authorship in educational settings (Ryan et al., 2025). As Brealant (2023) explains, infringement is assessed qualitatively based on substantial similarity rather than fixed numerical thresholds, yet the 30% figure endures as a practical guide for distinguishing derivative from original work. When educators revise more than 30% of AI-generated content in terms of narrative depth, ethical framing or instructional purpose, they move beyond facilitation and assume authorship. This percentage is not a legal rule, but a meaningful signal of human intervention, demonstrating epistemic authorship rather than mechanical output (Figure 1).

This model is visually captured as a Venn diagram illustrating the dynamic interplay between AI narrative generation and Human curation, which intersect to form the central zone of Collaborative iteration. Surrounding this core interaction is a larger, encompassing ring labeled Ethical accountability, symbolizing both a protective boundary and a critical filter that ensures thoughtful, responsible use of generative technologies in education. The diagram also includes directional annotations: one labeling the Educator as an Epistemic filter, highlighting the educator’s role in discerning and validating AI outputs, and another for Iterative feedback, showing the continuous loop of refinement between human insight and machine-generated content. This holistic visualization underscores that while AI can assist in narrative creation, it must always be guided by human judgment and ethical scrutiny.

This model helps reframe the educator’s role not as a passive overseer of machine output, but as an intentional curator of meaning. It offers a structured way to understand how AI can be used without surrendering pedagogical agency. And it reinforces that in the age of generative technologies, curation is authorship.

But this model also has limits and assumes:

• access to AI tools and the digital literacy to prompt effectively;

• institutional and cultural space for experimentation; and

• a commitment to reflective practice.

The Curated Authorship Model presumes that content may originate from AI, but meaning does not. Narrative development may be machine-generated but authorship emerges when a human educator intentionally develops and pedagogically shapes the content. This model, therefore, does not require human-first drafting, but it does require human-first judgment. The starting text can come from AI, but the starting meaning, intention and responsibility must come from the human. If the human simply accepts, copies or lightly edits AI text without pedagogical or ethical shaping, then it is not Curated authorship. It is prompted drafting or delegated authorship.

As my role evolved from author to curator, I found myself grappling with a foundational shift, not just how to use AI in education, but whether and why. In the absence of policy guidance or precedent, I turned inward, relying on reflective practice to navigate this terrain. What emerged was a heuristic, not a rulebook, but a compass. I call it the 4W’s Ethical Framework, a contextual reflection tool designed specifically for evaluating AI-generated content in educational settings.

Unlike general technology ethics tools, the 4W’s framework is tailored to the pedagogical, epistemic and ethical stakes of integrating generative AI into curriculum design. It differs from compliance rubrics in several ways (Table 4 and Figure 2).

This framework is particularly useful when AI meaningfully shapes content that enters the learning environment. It is not designed for incidental or low-stakes uses of AI. Educators should consider applying the 4W’s when AI is used to generate core educational materials such as case studies, syllabi or assessments, or when student engagement and learning outcomes are potentially impacted. Ethical considerations, including questions of representation, authorship and visibility, are also key indicators that the framework is warranted. Importantly, the framework is most relevant when the AI-generated content is used within teaching contexts, rather than for private drafting.

Conversely, the 4W’s need not be applied when AI is used for more mechanical or supportive tasks, such as grammar correction or syntax polishing. If the content is intended solely for personal use and not shared with students, or when AI is applied to purely technical outputs like data cleaning or visualization, the stakes are generally too low for the full framework. A rough heuristic might be that if less than 10% of the content is machine-generated, the 4W’s may not be necessary. Ultimately, this tool is designed to interrogate the generative and epistemic roles of AI in education, distinguishing it from general digital tools.

AI enables new kinds of educational innovation by enhancing generative capacity, increasing speed and providing structural scaffolding. For instance, AI can accelerate the early phases of designing a case study or allow instructors to dynamically model writing in the classroom. It enables the exploration of multiple narrative arcs or tonal variations. Yet, its value is not merely about efficiency; it unlocks a form of iterative creativity that was previously out of reach.

However, these gains come with trade-offs. AI can diminish important pedagogical values such as the development of individual voice, the productive ambiguity of early drafts and the recursive slowness associated with deep thinking. The fluency of AI-generated text can mask a loss of intellectual struggle. If AI is adopted too early in the creative process, it can displace the slow, uncertain work that fosters critical reasoning, a substitution that carries pedagogical risks.

Over-reliance on AI may lead to atrophy in essential cognitive and pedagogical practices. When students only encounter polished, AI-generated prose, they miss the “messy middle,” the false starts, revisions and uncertainty that constitute authentic intellectual labor. Similarly, educators who lean too heavily on AI may stop modeling the process of thinking through writing, which is often where the most meaningful learning occurs.

AI also introduces access to new voices and representational possibilities. It can generate multilingual and cross-cultural framings that expand the horizons of teaching. These perspectives can enrich educational content by surfacing narratives or angles that might otherwise be overlooked. However, this access must be critically curated. Without thoughtful engagement, AI-generated diversity risks becoming tokenistic or ethically problematic. Representation, after all, is not inherently meaningful unless it is handled with care.

Unlike traditional ethical codes or compliance-based rubrics, the 4W’s operates as a contextual reflection tool. It differs in three important ways:

1. It prioritizes reflection over compliance. The framework is not about checking boxes; it is about naming dilemmas as they arise.

2. It is contextual, not universal. These questions will surface different tensions depending on the educator’s discipline, audience, values and goals.

3. It is recursive, not linear. The questions are not asked once and filed away; they return at each stage of the creative cycle: ideation, drafting, revision and publishing.

Conceptually, the 4W’s draws inspiration from critical pedagogical ethics. Joan Tronto’s ethic of care reminds us that responsibility in education is ongoing, relational and grounded in attentiveness (Tronto, 1993). Helen Nissenbaum’s theory of contextual integrity alerts us that information technologies disrupt established norms of disclosure, intention and information flow, demanding that ethical judgments be situated within specific social contexts (Nissenbaum, 2004). The 4W’s borrows from both and is not meant to prescribe conduct, but to provoke intentionality (Table 5).

The 4W’s is not designed to solve ethical dilemmas. It is not prescriptive, nor is it intended to be exhaustive. It presumes a baseline of pedagogical literacy and ethical curiosity. It will not provide the “right answer” when the line between AI-generated and human-crafted becomes blurry. But it will ensure that the question is asked and that it’s asked in ways that foreground student experience, educator agency, and the evolving ethics of authorship. The framework works best in environments that prize process over product, and reflection over replication. It invites us to shift from reactive stances (“Is this allowed?”) to epistemic ones: What does this mean for how I teach, who I reach, and what I value?

In a world where AI can simulate thought, the educator’s role is no longer just to produce content but to curate its construction, disclose its scaffolding and reflect on its implications. The 4W’s Ethical Framework offers a map for that reflection not as a destination, but as a way to walk the terrain with more clarity and care in a world where in a world where intent no longer guarantees integrity.

The rise of generative AI in educational content creation is not just a technological shift. It is a transformation in pedagogical identity. It compels a rethinking of authorship, labor and epistemic authority in teaching. Having moved from traditional case writer to AI collaborator, I have experienced this shift firsthand, but the implications go far beyond my practice, they ripple across the broader educational ecosystem. Each stakeholder must now reconsider not just what we teach, but how we define the act of teaching itself.

Educators are no longer just authors; they are design curators. With AI now embedded in the creative process, the boundaries between composition and curation blur. The 4W’s Framework offers a starting point for reflecting on this transition. Whether applied during course planning, peer mentoring or syllabus development, the framework supports educators in making visible the often-invisible labor of ethical filtering, narrative re-shaping and learning design. Transparency becomes essential. If students are engaging with AI-assisted content, they deserve to know not to shame the process, but to understand it. Disclosure is not about disclaimers; it is about modeling intellectual responsibility in an age of simulated thought.

Moreover, educators need new vocabularies and structures for naming their labor. Prompting is not passive and iterating with AI is not automatic. The Curated Authorship Model reframes these practices as intentional, iterative and pedagogically consequential.

Higher education institutions must update their frameworks for evaluating teaching, authorship and academic integrity. Existing systems tend to treat content creation as a binary: either authored or plagiarized but AI introduces a third space: co-curation that falls outside this binary and demands policy innovation.

Key institutional actions might include:

• evolving authorship definitions to recognize iterative, AI-assisted labor as legitimate academic work;

• creating repositories of ethically curated AI learning materials, including prompt logs, revision notes and reflection journals; and

• building shared guidelines for transparency and attribution, especially in team teaching and interdisciplinary settings.

Students will inherit a world where AI co-authors much of what they read, write and analyze. Their educational experience must prepare them not only to consume such content, but to interrogate and co-create it.

This begins with AI literacy as epistemic literacy:

• Can students distinguish AI-generated text from human-written discourse?

• Can they identify where algorithmic flattening occurs in generic tone, simplified dilemmas or missing nuance?

• Can they critique a case not just for what it says, but for how and why it was constructed?

When I started down this rabbit hole of AI in education and, more specifically, AI in case writing, I came to a profound and unsettling realization. AI is poking holes in the educational boat that we all know and traditionally live within. Simply plugging those holes is what education is trying to do but we do not need to plug the holes because we need a new boat. I do not know what that new boat looks like or how different it will be from the current educational boat that we all know, but I do know that AI has poked too many holes in the current system.

When integrating tools like the 4W’s into classroom critique and modeling the Curated authorship process openly, educators can help students develop both critical distance and ethical agency when working with AI.

The first time I used generative AI to create a case study, I anticipated efficiency. What I did not anticipate was existential unease. The outline appeared in seconds, clear, structured and usable. Yet, as I reviewed it, I experienced a profound dislocation. I was no longer the author in the way I had previously understood that role. Instead, I had become something else: a narrative ethicist, a sense maker, a curator of content. That moment of discomfort catalyzed the genesis of this inquiry.

What began as a workflow experiment soon revealed itself as an epistemic shift. AI did not simply accelerate the process; it redefined my pedagogical identity. In response to this shift, I developed two conceptual tools: the Curated Authorship Model and the 4W’s Ethical Framework. These tools are not intended to provide definitive answers, but be reflective instruments designed to foster intentionality and ethical awareness in the face of accelerating technological mediation. They serve as mechanisms for slowing down when everything else is speeding up.

At their core, both tools return us to a foundational pedagogical question: If AI can simulate thought, what does it mean to teach in a world where thinking itself can be mimicked? The answer, I propose, lies not in rejecting AI outright or embracing it without critique, but in engaging with it intentionally. We must see ourselves not as passive implementers of algorithmic outputs, but as active curators of truth. Educators must shape, contextualize and disclose the knowledge they present. We are called to interrogate not only what we teach, but how it was constructed and for what purpose.

This reframing invites us to ask:

• What do I amplify?

• What do I revise?

• What do I omit, and why?

These are not merely technical decisions, they are epistemic and ethical in nature and they define the evolving responsibilities of pedagogy in the age of generative AI. Our role is no longer limited to the perception of the transmission of content. It now includes the careful curation of content’s origins, intentions and implications. In doing so, we teach not only what knowledge is, but also how it is created, by whom, and to what end.