Australian University Generative-AI Policy Analysis

Overview

A systematic, document-based comparison of how Australian universities regulate student use of generative AI. The project collects publicly available academic integrity policies, assessment procedures and student-facing guidance, and codes them against a common framework so that institutional approaches can be compared on like-for-like terms.

Problem

Australian universities have produced generative-AI rules at speed, but the resulting policies differ in scope, terminology, permissions and consequences. Students studying at different institutions — or even in different units at the same institution — face materially different rules for the same behaviour. Without systematic comparison, neither the sector nor researchers can say how consistent, complete or enforceable these policies actually are.

My role

Sole researcher: designed the study, built the sampling frame and coding framework, collected and coded all documents, and drafted the analysis. [Add supervisor/collaborator acknowledgements if applicable.]

Research questions

1. How do Australian universities define acceptable and unacceptable student use of generative AI?
2. What disclosure and acknowledgement requirements do policies impose?
3. How do policies allocate responsibility between institution, teaching staff and students?
4. Where do policies leave decision-relevant questions unanswered?

Data sources

• Publicly available academic integrity policies and procedures
• Assessment policies and unit-level guidance where public
• Student-facing generative-AI guidance pages
• Corpus spans five Australian universities; per-document counts will be published alongside the coding matrix once the analysis phase closes.

Method

Systematic policy and document analysis. Documents were identified through institutional policy libraries, captured with date stamps, and coded in a structured matrix. Coding decisions were documented with decision rules and revisited after a second pass. See the coding framework below.

Coding framework

Each policy was coded for:

• Scope — which students, tasks and tools the policy covers
• Permission model — banned / permitted-with-conditions / task-dependent / delegated to unit level
• Disclosure requirements — whether, how and where AI use must be acknowledged
• Authorship provisions — what the policy says about the student’s own contribution
• Verification expectations — whether students must check AI output
• Consequences — how breaches are classified and handled
• Clarity markers — defined terms, examples given, contact points

Findings

Findings are reported only once verified against the coded data set.

• [Add verified finding here]
• [Add verified finding here]
• [Add verified finding here]

Outputs

• Coded policy comparison matrix — [in preparation]
• Working paper — see the related publication entry on the Publications page
• [Add conference presentation or report details when confirmed]

Impact

• [Add verified impact here — e.g. citations, institutional use, media mention]

Limitations

• Analyses public documents only; internal guidance and actual enforcement practice are not observed.
• Policies change quickly; the corpus is a dated snapshot, not a live record.
• Coding involves interpretive judgement; decision rules are documented but a single-coder design limits reliability claims.

Lessons learned

• Terminology varies so widely between institutions that a shared coding vocabulary must be built before any comparison is meaningful.
• Date-stamping captured documents is essential — several policies changed during the study window.
• [Add further lessons as the project concludes]

Related publication

See [Placeholder] Manuscript under review on the Publications page.

Downloadable materials

• [Add coding framework PDF to public/downloads/ and list it in the frontmatter downloads field]