Faithful Writing Pedagogy in the Age of Generative AI: A Sabbath-Grounded Approach

Before the public launch of OpenAI’s ChatGPT in late 2022, discussions of AI in higher education were still relatively easy to avoid. While many people had begun to anticipate the impact of emerging AI technologies—some extolling the efficiencies promised by progressively sophisticated algorithms and others speculating apocalyptically about a world where these technologies gradually achieve sentience and deem humans unnecessary obstacles to artificial “superintelligence”—the sudden appearance of ChatGPT seemed to catch many educators off-guard. Even though previous iterations of OpenAI’s GPT technology and other Large-Language Models (LLMs) had been generating text for various uses, the ease and speed with which ChatGPT was able to generate text struck an ominous chord with journalists especially, who reported on the implications of the software for learning. Within a week after the launch of ChatGPT, The Atlantic bluntly declared, “The College Essay Is Dead.” ¹

One reason generative AI makes many in higher education uneasy is that the tool exposes some misplaced concerns about how students are taught and how learning is measured. For starters, programs like ChatGPT appear to replicate general subject matter knowledge on topics that have typically been the domain of disciplinary experts—though one does not need to spend much time with the tools to recognize how shallow that knowledge is. As experienced educators know, the challenge is not just delivering subject matter knowledge but equipping students through the right ordering of that knowledge, the process of shaping—or disciplining—that also includes the moral, ethical, and spiritual dimensions of teaching. Christian professors also hope that students will use the knowledge and skills they acquire for good ends as they make their own knowledge claims within the world, indicating the character development that is part of education at all levels. Generative AI now seems to offer students the ability to make knowledge claims without submitting to the work required to earn that knowledge and without developing the discipline that appropriately shapes it.

From the perspective of writing pedagogy, the troubling threat posed by AI is not that it will develop true experts in academic disciplines without guidance from credentialed educators but that it might allow students to generate polished text that gives them the appearance of expertise. This complicates the ability to assess student learning and to authentically credential graduates for future employers. The problem is intensified when writing is primarily treated as a measure of student learning rather than just one aspect of a complex process of disciplinary knowledge-making. Fortunately, while generative AI might expose certain risks in student writing, it did not create those risks. Even as the technology is new and some of the problems it poses may seem novel, many of the best practices that have emerged from writing studies can still provide a framework for the faithful and thoughtful integration of generative AI tools in our pedagogy, both in terms of how students are taught to use these tools and how they are encouraged not to use them.

Whether one views the capabilities of generative AI as dazzling or terrifying, the reports of the death of the college essay are greatly exaggerated. Generative AI programs may indeed be able to make short work of some typical college writing assignments. Authentic student writing, on the other hand, entails a complex set of performances that generative AI tools cannot replicate, although in some cases the software may be employed in certain parts of that process. This is not to say that writing pedagogy does not need to adapt to the presence of AI tools. Good writing pedagogy is always adapting as scholars continue to learn more about how students develop as writers and as new writing tools and technologies themselves develop, from word processors to editing programs and, now, generative AI. Christian educators ought to think critically about what these tools mean for how students learn to write and how students might be cultivated as people of virtuous character. The question is not “How do we protect our students from generative AI?” but rather “How do we, as Christian educators, respond to a world where generative AI is now part of the pedagogical equation?”

Christians are uniquely positioned to respond to generative AI without the fear or dismay accompanying the most apocalyptic forecasts of an uncertain future.² For Christians committed to the work of higher education, a faithful response to generative AI does not need to focus on creating new resources but, rather, on exploring existing resources that can help us to receive this offer of generative AI, integrate it, and incorporate it into the teaching process. By integration, I do not mean the wholesale acceptance of generative AI—that the goal of pedagogues now ought to shift to teaching students how to “engineer” more effective prompts into generative AI programs. Whether we want it or not, this technology has been offered to us. Whether we want it to be or not, it is also being offered to students. How will professors and students receive this offer together?³

While a fuller exploration of the ethical and theological implications of AI is beyond the scope of this paper, my commitment to the faithful integration of generative AI tools in writing pedagogy should neither be mistaken as an endorsement of those tools nor a passive resignation to their use.⁴ The challenges posed by AI are multi-disciplinary and extend to several areas of specialization. To be clear, I share many common concerns about AI’s impact on higher education. I am deeply skeptical of the false promise of reduced labor and the broad assumptions about the purposes and ends of work that AI is supposedly meant to liberate us from. I am also concerned about how faculty and administrators might turn to AI as a “more efficient” solution to some of the core tasks of higher education, such as recruiting students, designing syllabuses, responding to and assessing student work, and evaluating faculty performance. I am inclined to resist turning to AI tools, even as they have begun to appear in environments I use every day, including word processing software, my phone, email programs, and search engines. As much as I wish generative AI would go away, I am more interested in asking what it means to teach faithfully in the context I am in, not the context I wish I were in.

The answer, I will suggest, can be found in existing resources in the Church and in those that have emerged from the field of writing studies. Integrating these spiritual, theological, and pedagogical resources will allow committed educators to confront the challenges of generative AI without responding from a place of fear or anxiety. This paper offers one framework for what such an integration can look like. I will begin by addressing problematic assumptions about writing and writing pedagogy that have exacerbated fears about generative AI. I will then explore resources Christianity provides for helping us respond to AI. Specifically, the values of Sabbath rest and delight, as articulated by theologian and philosopher Norman Wirzba, can reorient a collective response to generative AI tools and inform a commitment to an apprenticeship approach to writing pedagogy, one that forms students as writers and practitioners of their disciplines and beyond.⁵ I will end by providing concrete suggestions for how faculty can faithfully integrate generative AI in their classes, teaching careful and responsible use of the technology when appropriate, and motivating students to avoid it when necessary.

Problematic Assumptions about Writing that Drive Fears of Generative AI

Some concerns about generative AI are rooted in existing frustrations about persistent problems that professors encounter in student writing. These frustrations are exacerbated by common assumptions about the relationship between writing and critical thinking and by an overemphasis on surface features of immature student writing. These assumptions, in turn, influence debates about the purpose and effectiveness of writing instruction, even as they are rooted in real disparities between what students are taught in first-year composition courses and the broad variety of ways that students are asked to produce writing in more advanced disciplinary courses and professional environments after college. The result is that generative AI tools now seem poised to exacerbate these longstanding concerns by undermining not only the role of writing in the college curriculum but also the other aspects of learning that student writing has traditionally been a primary means of demonstrating.

As Susan Peck Macdonald observed thirty years ago, composition scholars have largely abandoned an objectivist epistemology within which writing might be seen either as a transparent representation of thought or as “an all-purpose skill, an ability to translate thought with correctness and grace, but a skill unaffected by particular disciplinary problems or knowledge-making processes within specific disciplines.”⁶ Despite decades of research, the view that writing is a transparent and easily transferrable skill continues to inform popular and academic attitudes about writing-centered courses, which invariably come up short in eliminating student writing problems.⁷

First-Year Composition (FYC), especially, is often criticized for its failure to “fix” the problems that appear in student writing, which usually means errors of usage, mechanics, and grammar but also includes cognitively immature ways of approaching subjects, such as failures to develop clear and compelling theses, inappropriate uses of evidence, and difficulties with organization and structure. The fact that many FYC courses are housed in English departments encourages students and professors to regard the first-year writing course as “an isolated course, an end in itself, a general education requirement to be gotten out of the way.”⁸ A lack of disciplinary grounding in FYC can mean that students are often exposed to writing that does not serve any purpose beyond the classroom. Students in FYC might be asked to practice self-expressive genres or to explore a research topic that is interesting to them, but they frequently do not recognize the purpose of these assignments beyond satisfying the instructor’s subjective preferences. As a result, “writing papers is perceived by students as an activity to earn a grade rather than to communicate to an audience of readers in a given discourse community and papers are commodified into grades, grades into grade reports, grade reports into transcripts, etc. [. . . ] This condition also misleads students into thinking writing is a generic skill that, once learned, becomes a ‘one size fits all’ intellectual garb.”⁹ Despite being described as a foundational skill, college writing is often plagued by a kind of rootlessness as students and their professors struggle to integrate habits learned in the writing classroom into broader university life.

Perceptions about the shortcomings of writing instruction have led some educators to argue that generative AI has finally made FYC courses unnecessary since AI can “fix” many of the problems with student writing that professors in the disciplines tend to be most frustrated by. As Melissa Nichols asks earnestly, “Why do we need a required writing course if AI can do everything outside stakeholders want such a course to teach? Won’t it be easier to have AI take care of students’ biggest writing problems so that professors can focus on content? After all, these same professors often claim that they don’t have time to teach writing because they have too much content to teach.”¹⁰ Given the potential of generative AI to output text that lacks many of the mechanical errors made by student writers, it is not surprising that some students see the tools as an attractive solution to the frustrations of the context-free writing that they view as tangential to their larger educational goals—namely, earning the degree and finding a job. Likewise, as professors wrestle with the demands of reading and assessing student writing—especially writing that struggles to demonstrate coherence, expertise, or both—they may well see generative AI tools as an invitation to throw in the towel or drastically rethink the role of writing in their teaching. Meanwhile, emerging AI tools are capable of responding to student writing, thus promising to free professors from a time-consuming task that many experience as drudgery.

Of course, many educators understand that FYC is not a miraculous inoculation against conventional or disciplinary writing errors, and they recognize that the frustration, failure, and messiness of the writing process are formational for writers at all levels. For these educators, the entirely justifiable fear is that generative AI will lure students away from engaging in that challenging yet valuable process. As Corey Robin argues in the Chronicle of Higher Education, “Academic writing has never simply been about producing good papers. It’s about the ordering of one’s world, taking the confusion that confronts us and turning it into something intelligible, wresting coherence from chaos.”¹¹ This view is tied to the notion that writing, in Robin’s words, “pushes for a discovery of self” in a way that is similar to therapy, externally articulating an internalized self, making the self concrete, if only for a moment, from one draft to another, so writers can view their thoughts on the page and come to know both their subject and themselves better. This view has real merit, but it mistakes much about the purpose of writing in college. A too-narrow focus on writing as a path to self-discovery risks underemphasizing the fundamentally social and communicative aspects of writing, especially in academic contexts.

Some also argue that the writing process is humanizing, which entails that students who use generative AI tools may be participating in a collective process of dehumanization. According to Matthew Crawford, “If we accept that the challenge of articulating life in the first person, as it unfolds, is central to human beings, then to allow an AI to do this on our behalf suggests self-erasure of the human.”¹² For Christians like Crawford, these fears take on theological significance. Kevin Brown suggests, “The irreducible, non-transferrable essence of humans is that we are created in the image of God (Imago Dei). Part of realizing our humanity and enacting redemptive order is bound up in bearing out God-reflecting attributes such as creativity, productivity, moral reflection, and relational capacities.”¹³ While Brown recognizes some benefits that artificial intelligence might bring to healthcare, for example, he, like Robin and Crawford, worries about the consequences of using AI in writing situations. This is because “We often write to know. Thoughtful human expression through the act of writing does not only produce an outcome (a wedding vow, a sermon, an essay, etc.)—it is its own humanizing outcome.”¹⁴ While the process of writing involves many humanistic and theistic virtues, it is important not to overly romanticize the act of writing as a defining attribute of human worth. This line of thinking ought to be rejected in light of the theology of the imago Dei that considers the context of human disability.¹⁵ Describing the writing process as humanizing, as many writers are tempted to do, inevitably creates a hierarchy that excludes people from that humanizing process, including people with disabilities that make writing difficult or impossible; people faced with the challenges of writing in another language; and people who have not otherwise had access to writing pedagogies or the institutions where those pedagogies are produced or practiced. Indeed, for some, generative AI may be a means to participate in writing communities and may, therefore, give them access to forms of social belonging from which they have typically been excluded.

A Sabbath-Grounded Response to Generative AI Fears

As the previous section demonstrates, fears of generative AI reflect existing pedagogical and theological assumptions about writing. Before I consider how scholarship from writing studies helps to address these assumptions, I want to explore a theological framework that can help provide a more faithful response to fears about AI, both within and beyond educational contexts. In particular, the practice of Sabbath, as described by Norman Wirzba, offers a compelling framework in which Sabbath-keeping extends far beyond the observance of a regular “day of rest” to include delight in God’s ongoing, sovereign act of sustaining all creation. In this view, “Sabbath observance has the potential to reform and redirect all our ways of living.”¹⁶ In the context of concerns over AI, Sabbath observation reorients us to the story of creation, salvation, and eschatology in a way that reminds us that God is at the center of that story and we are not—nor is AI. This Sabbath grounding can reorder priorities around creativity, work, rest, and education—all of which are central concerns for writing pedagogy. Such a reordering prepares us to resist the lure of AI’s promise of a future free from work and tempers anxiety about an educational landscape grappling with novel AI technologies.

In Living the Sabbath, Wirzba challenges what he describes as a reductive tendency to think of “Sabbath observance as an add-on to the end of a busy week,” a time “to relax and let down our guard, to pause from the often anxious and competitive patterns of daily life.”¹⁷ The patterns of life can indeed be overwhelming. But in drawing from Abraham Heschel’s reading of the medieval rabbi Rashi’s midrash commentary, Wirzba invokes the notion that the Sabbath is not a mere break from work but is actually the culminating generative act of the creation story. As Wirzba writes, “After the six days of divine work creation was not yet complete. What it lacked, and thus what remained to be created, was menuha, the rest, tranquility, serenity, and peace of God.”¹⁸ Menuha, as described by Rashi, embodies “the sort of happiness and harmony that comes from things being as they ought to be,” and its formation “gave the whole of creation its meaning.”¹⁹ As “the climax of creation,” Sabbath is thus “the goal toward which all of our living should move.”²⁰

This teleological view reminds us that the Sabbath is not just something that happened at the beginning of time. The Genesis account is part of the story of Sabbath, but it is not the whole story. Sabbath is also rooted in the hope of Christian eschatology. The clearest indication of what Sabbath rest looks like is found in Jesus Christ, in whom creation, salvation, and eschatology intersect and who, according to St. Gregory, is “the true Sabbath.”²¹ From a cosmic perspective, this means the salvation that comes through Christ’s redemptive work is not focused narrowly on individuals but on the restoration of the whole of creation. As people of new creation, Christians share in this restoration not only by resisting the demands of empty busyness but by positively cultivating the practice of delight. In other words, as Wirzba writes, “Sabbath observance ought to and can be one of the church’s primary ways of witnessing to God’s emancipatory and peaceful ways in the midst of a violent culture that knows no rest.”²²

Recognizing Sabbath as the home of God’s menuha carries significant implications for reorienting our attitude toward work since work is typically understood as the opposite of rest. In contrast to the view of work as a form of divine punishment for sin, Wirzba argues that, “While it would be a mistake to collapse work and the Sabbath into one another—they are clearly not the same—it is no less an error to think that these two spheres or dimensions of life do not inform each other. Sabbath observance, if it is to be true and not a sham, must necessarily extend into our work lives.”²³ Our labor marketplace deems some types of work more meaningful than others, especially regarding the social status of different professions and how workers are compensated. AI now seems poised to replace many of these “less meaningful” types of work. A Sabbath framework corrects this perspective, for there is no type of work that is intrinsically more meaningful or more valuable than another. All may be occasions for worship since “human work finds its inspiration and fulfillment in God’s own work of healing, restoring, strengthening, and maintaining the life of creation. Our work, if it is to be good, must line up sympathetically with God’s.”²⁴ This includes the academic work of chemists, engineers, philosophers, and English professors, as much as it also includes the vocational work of plumbers, pilots, painters, park rangers, and pastors, and the nonvocational work of parenting, keeping a garden, serving members of our community, and loving our neighbors.

Viewing work through a Sabbath lens, refocused by the promised menuha of new creation, raises important questions about the work of educators and how students are trained to participate in the work of their disciplines—work which includes but is not limited to writing. For some, generative AI tools pose a threat to long-established means of credentialing students. Wirzba, however, rightly cautions against a vision of education that is more focused on credentialing students than it is concerned with the right formation of desire, and he calls for “guidance in the ways of authentic education, recognizing that without it we will not fully realize God’s loving intentions for us. We need clarity about the overall aims of formal and informal education so that together we can form communities that will lead us into God’s peace, justice, and joy.”²⁵This orientation aligns well with the missional commitments of Christian higher education. Still, Christian professors are just as susceptible to focusing on credentials and status, whether in how students are assessed, degrees are conferred, or institutional prestige is valued.²⁶

What is at stake is a question of goals, the ends to which educational efforts are directed. In contrast to an educative process that is focused on credentialing students—in order, perhaps, to make them more competitive candidates in a job market so that they earn more lucrative salaries or receive more prestigious recognition—Wirzba makes the case for Sabbath education as an apprenticeship whose goal is “a faithful and honest attunement to the world.”²⁷ Wirzba calls readers to “conceive education as our initiation into wholeness” and to think of education in ecological terms.²⁸ In contrast to a view of education as the transmission of knowledge, in which a magisterial professor initiates students into the secrets of their discipline, an apprenticeship model of education reframes the relationship as one of master practitioner and apprentice, with a focus on what it means to practice the craft of our disciplines: the ways of thinking and the forms of knowledge valued, the questions asked and the methods used to answer those questions, and the practices of communication—including writing—that learners develop as they work alongside and in tension with others in their disciplinary communities.

Wirzba’s notion of apprenticeship within Sabbath education is directed at reframing our relationship with the created world. This objective goes far beyond the local environments of higher education and, certainly, beyond writing classes. Yet, it would be a mistake to separate this more expansive view of apprenticeship from the everyday work of the university. Apprenticeship at any level lends itself to virtue formation since to be an apprentice is less a matter of acquiring the correct “information” about one’s craft but, instead of developing the habits and motions so that what is initially awkward or uncomfortable becomes fluid, automatic, and skilled.²⁹ When students (or their professors) see education, either on a small or large scale, as a series of tasks that will lead them to a specific degree so they can get the job, the result is disconnection and a kind of “rootlessness” which “leads to the dismemberment of the mind and of creation.”³⁰ By contrast, the Sabbath-grounded view of education described by Wirzba reconceives of “education as our initiation into wholeness.”³¹ As a carefully attuned and patient apprenticeship draws educators’ attention, and their students’, to the interconnectedness of God’s world, students are prepared less to be “knowers,” “workers,” or “employees” and more to be the kind of people who see “each other and ourselves as what we are: concrete expressions of the love of God.”³²

Many see AI tools (whether the tools that exist today or those that may be arriving) as promising a new sabbath that will be initiated through our technological prowess, a rest that humans will achieve when we have finally developed the tools that can complete our work for us. This prospect seems exciting for some, but the story of salvation makes clear that God’s menuha, much more than rest, will not be accomplished by AI or by any tool humans develop. This is not to say that God may not use AI to fulfill God’s eschatological promises, since God may use anything God wants—and often does so in surprising ways. AI is, in fact, part of God’s creation. Christians, though, have good reason to resist the pull of the overly utopian story that tells us that AI (or any other human technology) will give us rest by bringing about an end to work, just as they must resist the pull of the overly dystopian story that tells us that AI (or any other human technology) will prevent God from completing the work of new creation and ushering in the true menuha rest that Sabbath observance remembers and anticipates.

Writing Pedagogy and Apprenticeship

The concept of Sabbath, as described by Wirzba, invites us to reconsider our response to the notion of AI as a threatening offer. Through the notion of apprenticeship as an “initiation into wholeness,” Sabbath education provides a pathway for reconsidering the pedagogical response to new technologies, even as professors continue to address existing frustrations about student writing. If educators want their students to move beyond a credential-focused approach to writing, they must cultivate their classes as richer sites of development and apprenticeship. Yet, how would such a writing apprenticeship work? Even the best moments of writing instruction may be limited when students have difficulty transferring their writing to new classes and writing situations appropriately. These are problems that the discipline of writing studies has taken up over the past few decades. In this penultimate section, I turn to that broader field to show how guiding students through a more holistic picture of writing development can help them—and their professors—navigate the demands of academic writing, with and without generative AI.

The widespread availability of generative AI makes it ever more critical for educators to expand their pedagogical focus beyond single moments of instruction and consider the larger developmental picture of how writers progress from novice to expert writing. Summing up much of the research in writing studies, Anne Beaufort writes, “We know that writing is a complex cognitive and social activity and that the mental processes involved as well as the contextual knowledge bases that must be tapped are enormous. Writing skill is honed over a lifetime.”³³ This social and cognitive complexity indicates that simply outsourcing the more challenging aspects of writing development to generative AI, as some have cynically proposed, will still fall short of preparing students to be engaged critical thinkers and communicators within their academic disciplines and professional work. At the same time, many of the criticisms of FYC are justified since one or two courses cannot hope to supply students with the full range of writing skills they may be required to demonstrate in future academic and professional contexts. As I have already suggested, expecting FYC alone to accomplish this is a mistake. The tendency to do so greatly underestimates the need for later moments of instruction that carry on and reinforce pedagogical efforts to support writing development, as Chris Anson and several others have argued.³⁴

Within a developmental view of writing instruction, composition studies offers several useful guideposts for developing a faithful response to the most challenging problems of student writing in the age of generative AI. Richard Gibson and James Beitler have recently made the case for recognizing “the labor of writing as an aspect of Christian life, the life of discipleship.”³⁵ In their book Charitable Writing, Gibson and Beitler propose three foundational “spiritual threshold concepts” that can inform a Christian practice of writing: “humble listening, loving argument, and hopeful timekeeping.”³⁶ Rooted in the Christian virtues of humility, charity, patience, and hope, these threshold concepts demonstrate how many of the best practices in writing pedagogy can simultaneously stimulate growth in virtue. These concepts also complement other widely agreed-upon strategies, such as an emphasis on the social and rhetorical aspects of writing, the notion that writing is a knowledge-making activity, and the construction of disciplinary and professional identities through writing.³⁷

Effective writing pedagogy is about more than how students and their professors approach the writing process. It involves the general orientation of teaching itself. Alison Caviness Gibson offers a model of hospitable teaching in the first-year writing classroom, which is often a site of anxiety and vulnerability for students.³⁸ The widespread suspicion that students will use generative AI to circumvent serious academic work runs against this spirit of hospitality. A more hospitable writing pedagogy, on the other hand, helps professors meet student writers where they are—to assume that their writing efforts are offered in good faith, to expect that they will struggle in new writing situations, and to see the writing and feedback processes as an invitation to meet one another as people made in the image of Christ. This posture of hospitality informs any potential to cultivate the virtues of charitable writing in students.

When it comes to many of the high-stress areas for professors worried about inappropriate uses of AI, composition scholars agree that effective writing pedagogy cannot focus only on “form” (such as grammar and mechanics) or “content” (such as command of the material being written about). This bears emphasizing: successful academic writing does not just require a grasp of subject matter knowledge and the ability to write in fluent and compelling prose. Rather, it involves the development of many different kinds of knowledge that overlap with and inform each other. Anne Beaufort describes five such knowledge areas that expert writers draw from: discourse community knowledge, subject matter knowledge, rhetorical knowledge, genre knowledge, and writing process knowledge.³⁹ To these five, John Bean has also added a sixth area, information literacy, which is necessary for writers to engage carefully and critically with various research sources writers might encounter, including text generated by AI tools. According to Bean, the advanced disciplinary writing performances of expert, insider prose are “Often characterized by ‘eros’/transformation,” a claim meant to indicate “the importance of the affective domain as well as the cognitive in a student’s journey. The movement from outsider to insider status is often marked by a movement from surface learning to deep learning and by an awakening interest in the field’s questions and problems. Students start to discover their disciplinary ‘voices,’ their own agency as critical thinkers within a disciplinary conversation.”⁴⁰ As students move toward producing expert prose in the age of generative AI, professors can focus on strategic moments for developing students’ information literacy—teaching them, for example, how to approach generated text critically and skeptically—and developing students’ writing-process knowledge by teaching them to recognize the affordances and shortcomings of using generative AI within the wide variety of processes available to writers.

This picture of the multiple knowledge domains of expert academic writing offers a blueprint for how writing instruction can focus on key areas of tension related to generative AI tools. While generative AI might seem to lend itself to impersonal and generalized content generation, for example, Gibson and Beitler’s concept of humble listening in communities (both inside and outside the classroom) emphasizes the social aspects of writing; it provides an antidote to the view that student writing is primarily intended to display knowledge to a professor for evaluation, and it redirects students’ attention to a process that begins with generous listening to the real people who also become the audience for writing. Likewise, the principle of “loving argument” complicates the student belief that academic writing should either remain neutral or go on the offensive. Instead, Gibson and Beitler improve the metaphor of writing as a conversation (common in writing pedagogy) by encouraging students to see writing as a banquet. This metaphor invites students to think about the work that must be done before hosting a banquet, from sending out invitations to preparing the meal, and to consider what it means for writers to meet the banquet participants with hospitality. The value placed on writing as a socially symbiotic act can lower the stakes of writing and help students recognize the intrinsic worth of their efforts to communicate. In the process, students may also discern when generative AI tools are less suitable and sufficient to fulfilling the expectations of meaningful writing assignments.

In particular, Gibson and Beitler’s emphasis on hopeful timekeeping offers a powerful counterpoint to generative AI’s promise of a quicker and more efficient writing process unmarred by the anxieties of the blank page, the messiness of drafting, and the discomfort of uncertainty. Hopeful timekeeping resists what Gibson and Beitler call the “need for speed” and introduces students to slow writing as a positive good.⁴¹ Recognizing that many writers long for the kind of inspiration where writing flows almost automatically, Gibson and Beitler point to the Genesis creation accounts to highlight a connection between God’s creativity and ours. They argue, “When we create, though, we do not do so out of nothing (creatio ex nihilo [. . .]) as God can. We create out of existing matter (creatio ex materia).”⁴² In this, they draw attention away from the Romantic notion of writing as the product of inspiration or divine intervention and from the positivistic belief in writing as the unmediated transmission of thought. Consistent with Wirzba’s emphasis on Sabbath education, Gibson and Beitler show how the patient investment in the writing process—from reading and pre-writing to drafting, receiving feedback, and revising—reinforces the notion of writing as a craft that can be developed with practice and attention.

This developmental model of student writing, carried out by several professors working in cooperation to reinforce learning and develop new skills across a curriculum, is well poised to effectively address many of the problems that are endemic to an instrumentalist approach (such as plagiarism and failures in disciplinary thinking and critical complexity) and that generative AI would otherwise exacerbate. It is also worth noting that this model does not lend itself to a credential-focused approach to education. On the other hand, it aligns well with Wirzba’s Sabbath-grounded emphasis on education as an apprenticeship into wholeness. A significant trend of scholarship in writing studies has likewise focused on what Anne Beaufort calls “a social apprenticeship in writing.”⁴³ Like Wirzba, Beaufort suggests the need for instruction that moves beyond the artificial and rootless confines of the classroom and invites students into an apprenticeship within which they can practice a wide variety of authentic writing moments. This variety, moreover, can be reflected in the types of writing tasks that initiate students into disciplinary thinking as well as in student writing itself. If the role of writing in the curriculum is merely to measure what students have learned, to assess students’ understanding of subjects or their mastery of skills, these are writing tasks that generative AI can readily undertake on students’ behalf. Unsurprisingly, some students will take advantage of these tools, either as a study aid that might reinforce their learning and performance or as a shortcut around a risky process that involves struggle, discomfort, potential failure, and the threat of being denied the desired credentials.

On the other hand, when writing plays a purposeful role in a well-designed curriculum—when informal and formal writing tasks are connected to course objectives in a way that gives developing writers the chance to see the role that writing and communication play in the discourse communities they belong to—professors can be intentional about cultivating hospitable opportunities where students practice humble listening, loving argument, and hopeful timekeeping. Just as importantly, they can foster authentic opportunities for student learning. These opportunities focus less on surveilling students for possible violations of academic integrity and more on providing strategic moments to learn about the affordances and limitations of the various writing processes they engage in, including those that involve generative AI.

Practical Steps for the Faithful Integration of Generative AI Tools

Having considered critical theological and pedagogical resources for responding faithfully to the challenges of generative AI, I now turn in this final section to offer some practical guidance. While this advice is suitable for individual classes and assignments, it is meant to provoke broader conversation and collaboration between colleagues, within departments, and across programs, disciplines, and schools—as well as between academic and professional communities.

Clarify Language on Academic Integrity

One of the most frequently articulated objections to using generative AI tools in educational contexts is that such use constitutes cheating, a de facto violation of academic integrity akin to plagiarism or ghostwriting. While it is not difficult to see why many professors might jump to this assumption, it is not immediately obvious why all uses of generative AI tools necessarily constitute cheating, especially since similar smaller-scale writing support strategies—from accepting changes suggested by editing software or collaborating with a writing center tutor—are not typically seen in this way. As James Lang and others have argued, the definition of cheating itself is not universally agreed upon, and students’ values around originality and authenticity may differ from their professors’ values.⁴⁴

In a classroom informed by hospitable pedagogy, professors will avoid the cynical assumption that their students are searching for ways to cheat and will be transparent about when and how the use of outside resources goes beyond what is acceptable. As students struggle with complex course concepts or challenging texts, they might turn to generative AI for a summary or an explanation that supports their efforts to understand those concepts better, just as prior generations of students have used resources like Wikipedia, YouTube, online study guides, or CliffsNotes. Few consider turning to these learning supplements as a violation of academic integrity. Regardless of the resource, students need to develop the information literacy that leads them to approach all study aids with a high degree of skepticism and critical thinking. Professors also articulate the value of students puzzling through challenging problems by themselves and with the support of others (including classmates, study tutors, writing center coaches, and professors). Rather than labeling all uses of generative AI with the charged moral language of cheating, a better approach to generative AI tools will teach students to be skeptical of generated text and will help them see how its outputs fall short of the more complex work they are capable of doing.

Don’t Try to Ban the Tools

Blanket prohibitions against all uses of generative AI are likely doomed to fail, especially if they rest on the assumption that students and professors agree that using the tools constitutes cheating. In fact, not all professors and professional environments hold the same views. While some professors may ban the tools outright, others may permit students to use them or even require them to do so, especially as AI capabilities are built into common programs people already use. Those who allow students to use generative AI tools can have as much, if not more, pedagogical rationale as those who prohibit them, and vice versa. This creates a complicated situation for students who must navigate the expectations of different professors. From a rhetorical perspective, students might assume that the competing rationales they encounter simply reflect individual preferences rather than well-considered pedagogical and disciplinary motives.

Resisting the urge to ban generative AI tools outright does not have to mean students are given free rein to outsource their writing to those tools. As professors cultivate practices of humble listening in their classroom communities, there may be circumstances where a ban on generative AI is appropriate, such as in online discussion boards that serve as a proxy for or extension of in-class discussions. The use of generative AI in these environments is best understood as instances of impersonation. When students submit generated text to an online discussion board, they violate the expectation that participants are interacting with human presence and agency. Professors can address these inappropriate applications by introducing classroom instruction on presence and community that clarifies the social and interpersonal function of this kind of writing.

It may also be appropriate to ask students not to use generative AI in specific situations, such as when an assignment is meant to help students work through the messiness of a challenging issue in writing. There is real value that comes from working through ill-formed problems in writing. When an assignment is tied to specific learning objectives, it is important to let students know the objective of that task in light of the learning outcomes for the course. For example, an assignment that asks students to summarize and respond to a challenging argument might include instructions for students to avoid outside resources like Google, Wikipedia, and ChatGPT. In support of focused learning objectives, such an assignment will also assure students that they are not expected to understand or write about everything included in that argument. To cultivate the spirit of humble listening, students can be asked to share their responses in a follow-up class where they can help each other fill in gaps in their understanding. Or, if the assignment also makes use of informal writing, assessment is best concentrated on students’ efforts to negotiate the demands of the assignment, which means making room for messy prose and usage errors that are characteristic of challenging cognitive work.

Don’t Overcompensate with More In-Class Writing

In the aftermath of the public release of ChatGPT, many educators have proposed a shift to more in-class work. In recognizing the futility of attempting to ban the tools, some have suggested that the best solution is to create an assessment environment where students will no longer have access to electronic tools. Notwithstanding the concerns many have about generative AI, the quickness with which some have embraced this position is surprising, especially given the dramatic shifts in higher education that have included the broad adoption of online learning management systems and hybrid teaching formats.

While there are undoubtedly good pedagogical reasons for requiring in-class writing in strategic moments, an overemphasis on in-class writing reinforces the view that education is primarily concerned with credentialing students rather than initiating them into an apprenticeship of the work scholars do in their disciplines. To begin with, in-class writing undermines the crucial role that the writing process can play in exploring challenging problems, struggling with complex texts or concepts, and moving toward authentic academic discovery. A policy of in-class writing may also produce disparate effects for many students, raising questions about accessibility and learning accommodations and challenging efforts to cultivate the classroom as a genuinely hospitable space. At many universities, some students with learning disabilities are granted time-and-a-half on timed assignments, while others have accommodations that permit them to take exams on computers, regardless of a professor’s technology policies. These accommodations are an important part of how schools fulfill their mission of serving students. The impulse toward mandatory, in-class, handwritten exams risks undermining those values.

On the other hand, in the spirit of an apprenticeship model of education, it is worth admitting that student writing has sometimes played an outsized role in measuring learning. If generative AI pushes some educators toward alternative forms of assessment that decenter the role writing has played in classes, moving from traditional term papers and essay exams to in-class presentations, multi-media projects, and collaborative assignments, this move is worth celebrating. Despite its long history, the traditional term paper is not always the best tool for fostering or assessing learning, and the shift to include other, more authentic forms of assessment is long overdue for some. While these alternative forms of assessment decenter the role that writing has played in classes, they also rightly encourage students to think more broadly about different ways of knowing that are characteristic of their disciplinary communities and prepare students to conceptualize and present their learning in ways that more closely match authentic work done by professionals within their disciplines.

Teach Openness and Transparency

Regardless of the policies about generative AI tools set by a professor, program, or school, it is important to cultivate an environment of openness, transparency, and trust rather than suspicion and surveillance. This includes creating legitimate pathways for students to cite or otherwise acknowledge their use of generative AI tools without facing severe penalties. Most students come to college classrooms having been warned that there are dire consequences for any violation of academic integrity, and an entire industry of text-matching programs has grown to help instructors identify parts of student writing that may be plagiarized. They are now being tuned to identify writing produced with generative AI tools. These programs may be useful for surveilling students but not for cultivating strong writing habits. Students, of course, can also access many of these programs. In instances of possible plagiarism, students can be alerted to places in a paper (whether truly plagiarized or entirely original) that might be red-flagged and consequently adjust those parts of their writing to pass the detection software.

In the same way, a small cottage industry of software tools, many free online, promises to “humanize” AI-generated text to bypass detection software. Most of the software deployed in this game of cat and mouse, whether from big industry leaders or open-source startups, includes terms of service that grant the tool access to student writing, which it may then add to the tool’s dataset. Students usually have little choice but to consent to these terms. While serious violations of academic integrity do occur, an overreliance on these tools runs counter to the habits most professors want to develop in their students, not only because it can become a self-fulfilling prophecy but because it outsources crucial aspects of professors’ engagement with student work to for-profit companies that cannot care for students.

Within a developmental view of student writing, it is to be expected that students will struggle in some places. In many instances, students are still learning conventions for appropriate academic citation. As they move between courses, disciplines, and rhetorical environments, students may have to navigate different citation conventions and relearn how and when to use (or avoid) generative AI tools. Professors cannot expect another class to have addressed all the concerns that might motivate the conventions of a new disciplinary community. Most of the major style guides have developed standards for citing AI-generated texts, and students need to be trained in these citation conventions. If the unauthorized use of generative AI tools is only treated as cheating and dishonesty, educators will miss out on opportunities to redirect students’ attention toward better practices of transparency that are expected of scholars in their disciplinary communities. On the other hand, when these moments are treated as occasions to guide students in the disciplinary and ethical imperatives to provide appropriate attribution, students begin to develop agency as writers, even as they learn to engage more critically with generative AI tools.

This practice of openness starts with professors’ transparency about their own practices. When educators draw from other resources in their course planning, whether adapting another colleague’s lesson plan, supplementing their knowledge of course material with outside reading, or even brushing up on a rusty topic by reviewing a Wikipedia summary, how often do they lay bare they own learning and preparation process for students and give attribution to those resources? If they use generative AI software to produce a new image for a slide deck, or if they pull pictures or diagrams from the Internet, do they provide appropriate source information? When professors name their practices for students, they model their vulnerability as continual learners whose expertise and teaching depend on a larger ecosystem, who still discover gaps in their knowledge that need to be filled, and even as people who feel the push of a deadline in low-stakes situations that sometimes might lead them to the most quickly available, rather than the best, resources. To be sure, there is an important difference between the status of credentialed professors and students who are at an earlier stage in their development as learners. Vulnerability like this can seem risky when the status hierarchy between professors and students must be preserved at all costs. Yet hospitable pedagogy is willing to confront the risk that comes with troubling that hierarchy. If professors are focused on guiding students as apprentice learners, there is a real value to demystifying their own processes.

Continue to Emphasize the Process over the Product

As I have mentioned, many educators are concerned that generative AI tools will be used as a shortcut around the messy process of writing, which not only includes moving from rough drafts to polished prose but also the riskier acts of formulating arguments and making claims, working through confusion, navigating anxiety and uncertainty, and seeking out critical feedback. As Gibson and Beitler remind us, this process requires hope.⁴⁵ It also requires resilience and vulnerability, and engaging in it can help develop a writer’s resilience and willingness to be vulnerable. However, such intrinsic goals may not be as motivating for students who are more concerned about the grade they will receive on the final product than the promise that writing will build character or achieve self-discovery. When professors are invested in the possibilities for the writing process, they can continue to foster opportunities that prompt students to engage in that process meaningfully.

According to Kathleen Blake Yancey, “As our knowledge about writing has deepened, we have understood that composing processes also vary according to at least three factors—the individual writer, the genre being composed, and the rhetorical situation.”⁴⁶ In other words, writing processes vary between individuals and across disciplines. Professors can build on students’ understanding of writing processes by intentionally incorporating the writing process into their assignments. The writing process is already a key emphasis of most FYC programs, enacted primarily through assignments that go through a series of required drafts that receive feedback from peers, writing center tutors, and professors. Many writing programs have also adopted a portfolio or ePortfolio grading model, which gives students the freedom to revise selected pieces of writing and present their work alongside reflections that provide insight into the students’ writing processes. Rather than expecting an FYC course to give students everything they need to know about the writing process, professors can reinforce students’ ability to navigate it in subsequent writing situations.

In an apprenticeship model of education, this begins with professors talking to students about their processes and allowing students to explore the different approaches to writing that are most effective for them. Professors can also take several steps that can reduce the need for students to turn to generative AI tools or other shortcuts that might undermine the learning goals for the course. Many faculty already incorporate some of these steps into their classes. This is a reminder that one of the best ways to respond to generative AI tools is by leaning into existing strengths. For example, a scaffolding approach to assignment design will divide a formal writing assignment into several smaller steps with lower stakes.⁴⁷ Low-stakes assignments can also include revision components that allow students to develop flexible thinking around a topic, encounter more complexity and a more comprehensive range of perspectives, and try out different arguments along the way to producing a stronger finished paper. While this may sound like an invitation to increase faculty workload by creating more assignments to grade, most low-stakes assignments do not require the same detailed feedback given to more formal writing.⁴⁸

As educators encourage students to invest in the writing process, they can also turn students’ attention to the social aspect of writing. Many students believe writing is mainly meant to be a solitary endeavor and see collaboration as out of bounds. This is another reason why generative AI tools appeal to many writers. They allow a writer to interact with a chatbot on their terms while maintaining the appearance of an individual writing performance. To counter this misconception, students can be directed toward human resources at several steps in the writing process through in-class group work, well-structured peer review workshops, one-on-one conferences with the professor, and visits to the university writing center. When tied to a scaffolded sequence and various low-stakes assignments, this social element can create opportunities for students to share works-in-progress with others and help them feel supported as part of a hospitable writing community.

Conclusion

A mutually supportive, hospitable writing community may be one of the best antidotes to generative AI’s offer of efficient writing as a solitary act. That does not mean that the tools will not be overwhelming to professors who are devoted to the patient and careful investment in students as people of character and experts in their disciplines. The suggestions offered here reflect my commitment to a Sabbath-grounded perspective, rooted in the desire to let God’s menuha take precedence over my own anxieties about a changing educational landscape. They are also informed by my commitment to fostering a developmental view of writing, recognizing that each student’s apprenticeship toward writing expertise will vary depending on their strengths and the expectations of the disciplines and genres they practice. Both commitments challenge the notion of writing and writing instruction as an individual endeavor. As educators work alongside each other and their students to negotiate the demands of academic writing and work in an age of generative AI, may the immediate pedagogical and ethical dilemmas posed by AI tools be reframed faithfully in light of God’s bigger story.

Cite this article

Traynor Hansen, “Faithful Writing Pedagogy in the Age of Generative AI: A Sabbath-Grounded Approach”, Christian Scholar’s Review, 54:2 , 55-76

Footnotes

Stephen Marche, “The College Essay Is Dead,” The Atlantic, December 6, 2022, https://www.theatlantic.com/technology/archive/2022/12/chatgpt-ai-writing-college-student-essays/672371/.
For an example of how Christian doctrine can inform our views of AI, see Michael D. Langford, “A Theological Framework for Reflection on Artificial Intelligence,” in AI, Faith, and the Future: An Interdisciplinary Approach, eds. Michael J. Paulus, Jr. and Michael D. Langford (Pickwick Publications, 2022), 70–94.
In using the language of the “offer,” I am drawing from Samuel Wells’ work in theological ethics. Wells turns to theatrical improvisation to describe how Christian practices can prepare us to respond faithfully to threatening or promising situations. Each “offer,” in improvisation, can be “blocked” or “accepted.” Wells advocates for “overaccepting,” which he describes as “accepting in light of a larger story,” i.e., the story of Scripture. See Samuel Wells, Improvisation: The Drama of Christian Ethics (Baker Academic, 2018), 109.
For more on ethical problems associated with generative AI and LLMs, see Emily Bender and Timnit Gebru et al., “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜” in Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, March 1, 2021, https://doi.org/10.1145/3442188.3445922; Shannon Vallor, The AI Mirror: How to Reclaim Our Humanity in an Age of Machine Thinking (Oxford University Press, 2024). For several interdisciplinary Christian perspectives, see AI, Faith, and the Future: An Interdisciplinary Approach, eds. Michael J. Paulus, Jr. and Michael D. Langford (Pickwick Publications, 2022).
See Norman Wirzba, Living the Sabbath: Discovering the Rhythms of Rest and Delight (Brazos Press, 2006).
Susan Peck Macdonald, Professional Writing in the Humanities and Social Sciences (Southern Illinois University Press, 1994), 5.
For a recent overview of the state of scholarship in writing studies, see Naming What We Know: Threshold Concepts of Writing, eds. Linda Adler-Kassner and Elizabeth Wardle (Utah State University Press, 2015).
Anne Beaufort, College Writing and Beyond: A New Framework for University Writing Instruction (Utah State University Press, 2007), 9.
Beaufort, College Writing and Beyond, 10.
Melissa Nichols, “Eliminate the Required First-Year Writing Course,” Inside Higher Ed, November 14, 2023, https://www.insidehighered.com/opinion/views/2023/11/14/eliminate-required-first-year-writing-course-opinion. The usefulness of FYC has been debated both within and outside of writing studies. See also Sharon Crowley, Composition in the University: Historical and Polemical Essays (University of Pittsburgh Press, 1998).
Corey Robin, “The End of the Take-Home Essay?” The Chronicle of Higher Education, August 24, 2023, https://www.chronicle.com/article/the-end-of-the-take-home-essay.
Matthew Crawford, “AI as Self-Erasure,” The Hedgehog Review, June 11, 2024, https://hedgehogreview.com/web-features/thr/posts/ai-as-self-erasure.
Kevin Brown, “The Threat of AI Is Not That It Will Consciously Take Over: The Threat Is That We Will Unconsciously Let It,” Blog, Christian Scholar’s Review, November 28, 2023, https://christianscholars.com/the-threat-of-ai-is-not-that-it-will-consciously-take-over-the-threat-is-that-we-will-unconsciously-let-it/.
Brown, “The Threat of AI Is Not That It Will Consciously Take Over.”
See Langford, “A Theological Framework for Reflection on Artificial Intelligence,” 79.
Wirzba, Living the Sabbath, 14.
Wirzba, Living the Sabbath, 23.
Wirzba, Living the Sabbath, 33.
Wirzba,Living the Sabbath, 33.
Wirzba, Living the Sabbath, 33.
Wirzba, Living the Sabbath, 43.
Wirzba, Living the Sabbath, 52.
Wirzba, Living the Sabbath, 91.
Wirzba, Living the Sabbath, 95.
Wirzba, Living the Sabbath, 131.
See also Jeffrey Bilbro, “Finding a (Real) Christian College,” Christianity Today, May 1, 2024, https://www.christianitytoday.com/ct/2024/may-web-only/finding-real-christian-college-formational-education.html.
Wirzba, Living the Sabbath, 132.
Wirzba, Living the Sabbath, 135.
The language of apprenticeship and craft is central to Henry Staten’s critique of the Romantic concept of art and the idea of the “genius creator,” both of which remain influential in popular and Christian assumptions about writing. These ideas also tend to be problematic for writing pedagogy. As an alternative, Staten describes the development of techne-craft as a kind of cunning, or know-how: “From the techne standpoint, the mystery, and the fascination, is primarily in the cunning by means of which Art and all other systematic, goal-oriented, knowledge-based forms of human activity are carried out—the cunning that organizes materials, methods and the artisans’ actions in the most effective way.” See Henry Staten, Techne Theory: A New Language for Art (Bloomsbury Academic, 2019), 5.
Wirzba, Living the Sabbath, 134.
Wirzba, Living the Sabbath, 135.
Wirzba, Living the Sabbath, 141.
Beaufort, College Writing and Beyond, 6. Scholars working within the fields of writing studies and novice/expert theory have offered several models for conceptualizing an emerging writer’s development. Susan Peck Macdonald offers a four-stage continuum through which writing expertise is developed, and Patricia Alexander outlines a three-stage model that begins with acclimation, progresses to competence, and finally moves to proficiency/expertise. Whether conceptualized as three or four stages, these models helpfully trouble the binary distinction between novice (student) and expert (professional) writing and focus to the middle space, where pedagogical energy can be directed. See MacDonald, Professional Writing in the Humanities and Social Sciences, 187; Patricia A. Alexander, “The Development of Expertise: The Journey from Acclimation to Proficiency,” Educational Researcher 32, no. 8 (November 2003).
See Chris M. Anson, “Crossing Thresholds: What’s to Know about Writing Across the Curriculum,” in Naming What We Know, eds. Linda Adler-Kassner and Elizabeth Wardle (Utah State University Press, 2015), 201–219.
Richard Hughes Gibson and James Edward Beitler III, Charitable Writing: Cultivating Virtue Through Our Words (IVP Academic, 2020), 1.
Gibson and Beitler, Charitable Writing, 13.
See Adler-Kassner and Wardle, Naming What We Know.
See Alison Caviness Gibson, “Welcoming the Student Writing: Hospitable Christian Pedagogy for First-Year Writing,” Christian Scholar’s Review 52, no. 1 (Fall 2022): 71–89.
Beaufort, College Writing and Beyond, 19.
John C. Bean and Dan Melzer, Engaging Ideas: The Professor’s Guide to Integrating Writing, Critical Thinking, and Active Learning in the Classroom, 3rd ed. (Jossey-Bass, 2021), 220–21.
Gibson and Beitler, Charitable Writing, 139–140.
Gibson and Beitler, Charitable Writing, 146.
Anne Beaufort, “Learning the Trade: A Social Apprenticeship Model for Gaining Writing Expertise,” Written Communication 17, no. 2 (April 2000): 214.
James M. Lang, Cheating Lessons: Learning from Academic Dishonesty (Harvard University Press, 2013), 165.
Gibson and Beitler, Charitable Writing, 113.
Kathleen Blake Yancey, “Writers’ Histories, Processes, and Identities Vary,” in Naming What We Know, eds. Linda Adler-Kassner and Elizabeth Wardle (Utah State University Press, 2015), 52.
Research on academic integrity suggests that lowering the stakes for assignments can also de-incentivize students to take shortcuts. See Lang, Cheating Lessons, 105.
For professors looking for more practical guidance on scaffolding writing assignments and developing effective strategies for offering feedback on student writing, see Bean and Melzer, Engaging Ideas.