Making the student writing process transparent: Version control as a practical & pedagogical response to LLM enabled academic misconduct

Haga clic aquí para la traducción al español (Google Translate)

Introduction: From policing to process

Since the release of ChatGPT to the general public in late 2022, large language models (LLMs) have progressively blurred the lines between authentic student writing & ‘LLM assisted’ writing. Such writing is not entirely the student’s own work & does not demonstrate their independent knowledge, skills, & attitudes. This has led many pundits to declare an end to the traditional academic assessment model & perhaps the devaluation of higher education. In response, many institutions have relied on detection software & stricter integrity policies, but these measures often foster mistrust & require expensive & intrusive surveillance.

A more constructive approach is to make the writing process transparent rather than engaging in an ‘arms race’ of plagiarism detection & anti-cheating measures vs. student ingenuity & 3rd party plagiarism evasion tools. Version control systems, the automatic recording of edits & revisions integrated into many online word processors, reveal how a text evolves, turning writing into a visible process instead of an opaque product. This transparency discourages misconduct while supporting learning, allowing teachers to see authentic drafting & reflection rather than just a polished final text.

This article argues that version control offers not only a practical response to LLM enabled misconduct but also a pedagogical opportunity: it moves the focus from detection & punishment to the student writing process, thereby aligning academic integrity with opportunities to cultivate genuine learning.

The pedagogical value of transparency

Version control systems are already embedded in common authoring software such as Google Docs, Collabora (the online collaborative implementation of LibreOffice, free & open source), the free & open-source platform MediaWiki (as used by Wikipedia), or even Git (a highly flexible & well-supported version control system for software development). They record every edit, revision, & comment made within a document. Each change is automatically timestamped, creating a detailed history of how a text evolves. This visibility transforms writing from a mysterious, finished product into a transparent process. Teachers can see the recursive drafting, reorganising, rewording, & refining that characterise authentic academic writing. Conversely, sections of a polished essay that appear in a single edits become an unmistakable signal of inauthentic authorship.

For most students, simply the knowledge that their writing process is observable is enough to discourage misconduct. Knowing that teachers can trace when & how their work was produced promotes a culture of openness & reflection. Importantly, this transparency does not rely on invasive proctoring software, ‘locked-down’ laptop computers or tablets, or surveillance; version control is built into the normal act of writing & is also typically visible to the authoring student as a useful tool for the writing & revision process.

Writing as a process, not a product

The educational benefits of version control extend well beyond deterring misconduct. It also enables process-based assessment, allowing teachers to evaluate how students plan, draft, edit, & respond to feedback. The focus can shift from grading the final text to assessing the intellectual journey that led there, an approach that aligns with well-documented models of the cognitive processes of academic writing (e.g. See: Chan, 2018).

Contrary to many students’ beliefs, authentic proficient academic writing is inherently iterative & recursive. Version control can reveal that writers frequently review quotations, rephrase source material, check coherence & organisation, & reorient their argument as their ideas mature & these revisions are rarely linear or uniform. In the educational context, their traces reveal far more about a student’s understanding than an essay submitted as a file ever could. Thus, version control systems are a valuable tool for making these processes visible allowing teachers to diagnose where difficulties arise & to provide targeted pre-instruction, transforming version control into a powerful tool for formative assessment & opening opportunities to support ‘learning to write’ & ‘writing to learn.’

Detecting inauthentic patterns

The history of edits recorded by version control systems further support this approach. Teachers can observe how a document develops over time & recognise unnatural work patterns, such as an entire essay produced within minutes or no activity between drafting deadlines. These anomalies offer strong indicators of potential misconduct. They also make it far more difficult for students to submit ‘LLM dump-and-submit’ essays that lack any sign of incremental development.

Moreover, the same systems can reveal the nature of collaboration & authorship. Metadata shows who contributed what, when, & from where, providing transparency around group work & deterring unauthorised collaboration or ghostwriting. Comment histories can show whether genuine peer feedback occurred or whether the text appeared fully formed, rather than by careful consideration of student review & subsequent development of ideas & language.

Encouraging metacognitive reflection

An additional pedagogical advantage emerges when teachers invite students to reflect on their revisions. Simple prompts such as “Why did you change this paragraph?” or “What were you trying to improve here?” encourage metacognitive & metalinguistic awareness or the writing & revision process. Such reflection can be built into grading criteria, rewarding students not only for what they write but for how they engage with revision & feedback. Importantly, it is nearly impossible for an LLM to generated a text with a convincing fabricated record of reflective revision. This automated, frictionless ‘show your working’ approach helps both teachers & students alike.

Toward a culture of process accountability

Embedding version control into assignment design represents a shift in academic culture. Rather than framing integrity as a matter of surveillance & punitive measures, this approach integrates transparency & accountability directly into the writing workflow (much like scientific research practices are moving towards more transparent models with pre-registration & open data). Students understand that their teachers can see their writing process, but they also come to see revision as part of learning rather than an afterthought. This transparency helps demystify writing as a process to be guided & supported, & cultivates trust between students & teachers, a more productive approach than adversarial, punitive policing.

For institutions, the benefits are equally compelling. Version control provides a fair & transparent audit trail for integrity investigations. In cases of dispute, educators can produce objective evidence of a document’s evolution, eliminating the need to rely solely on LLM detection algorithms that are often inaccurate & biased against non-native speaker students, & can be easily defeated by LLM ‘humaniser’ software.

Anticipating the arms race

Sceptics will argue that, inevitably, students & developers will create tools to simulate these processes, LLM web browser plugins that artificially ‘replay’ student-like drafting behaviours. Such a future is theoretically possible but practically distant. Generative systems require vast, domain specific datasets to mimic authentic student writing patterns, & unless hosts of document platforms voluntarily release this data, the development of convincing forgery generators will remain improbable for years, if not decades. Moreover, version control operates at the server level, integrated with the document management infrastructure itself, rather than through superficial plugins.

From misconduct prevention to pedagogical enrichment

What begins as a practical safeguard against LLM enabled misconduct quickly reveals itself as a powerful pedagogical tool. Version control systems allow educators to observe & support the cognitive & linguistic development of their students. It transforms writing into a transparent, teachable process & encourages authentic academic writing habits.

Far from only being a deterrent, version controlled writing environments could support cultivating proficient & productive writing habits, aligning integrity with pedagogy & assessment. By embedding transparency into the architecture of academic work, institutions can protect authenticity & integrity while simultaneously deepening students’ understanding of what it means to learn to write & write to learn.

References

  • Chan, S. (2018). Defining Integrated Reading-into-Writing Constructs: Evidence at the B2–C1 Interface. Cambridge University Press.
  • Miller, B. (2023). What’s the Diff? Version History & Revision Reflections. Middle Tennessee State University (MTSU). https://mtsu.pressbooks.pub/engl1020/chapter/version-history/

Appendix

To further illustrate how version control systems work in practice, the following video demonstrates some of the features of the version control system in Collabora, a fully GDPR compliant, free & open source online shared document editing & collaboration platform:

Summary of Collabora Online (LibreOffice-based)

  • Collaboration: Supports real-time editing & version control.
  • Ease of use: Familiar LibreOffice-style interface, intuitive for students.
  • Offline access: Available via the Collabora Office desktop app.
  • Integration: Integrates well with Moodle, Nextcloud, ownCloud, & other open-source LMS tools (Moodle plugin repository: https://moodle.org/plugins/mod_collabora).
  • Accessibility: Includes keyboard shortcuts & screen reader support.
  • Compatibility: Excellent support for Microsoft Office formats. Long-term commitment to Moodle integration.
  • Privacy & compliance: Secure & fully GDPR compliant; no external data sharing; no ads or tracking.
  • Hosting: Self-hosted on school or government servers for full data ownership.
  • Licensing: Free & open-source software with optional paid support.
  • Best for: Universities, colleges, & institutions prioritising privacy, open-source tools, & data control.
  • Website: https://www.collaboraonline.com/

Disclaimer: I have no affiliation with Collabora or LibreOffice.