UQ’s rules for using AI
As educators, we have an opportunity to guide our students in using AI responsibly, developing critical judgement and upholding academic integrity — foundations that echo the UQ values of truth, excellence, creativity, courage, integrity, respect and inclusivity.
UQ’s Responsible AI Framework (PDF, 87.4 KB) supports informed, context-sensitive decisions about when AI should be used, how it should be used and when it should not be used.
Staff and students share responsibility for meeting UQ’s academic integrity, conduct, privacy, intellectual property, copyright and other relevant obligations when using AI. Educators should make assessment conditions clear, model responsible practice and support students to understand how to maintain their responsibilities, voice and exepertise.
Students may use AI tools responsibly in their study and open assessment tasks. Supervised 'secure' assessments may restrict AI use.
UQ’s approach to AI use and academic integrity expectations
"The ongoing advances in Generative AI technologies present both opportunities and challenges for teaching and assessment. While protecting academic integrity is critical, we also acknowledge the increasing prevalence of AI in our everyday lives, making it critical for us to help students to understand its ethical and effective use.”
- Professor Kris Ryan, University of Queensland Deputy Vice-Chancellor (Academic)
Assessment tasks are classified as secure or open, with matching automated text added to course profiles to improve consistency and clarity. Task instructions must still specify the materials and tools students may use and any disclosure requirements.
- Secure: assessment undertaken by identified students under supervision, providing evidence of their achievement of the learning outcomes and confidence that assessment outcomes reflect students’ own learning. Secure assessment instructions should identify the materials and tools students may use when completing the task, including whether AI use is allowed.
- Open: assessment completed without supervision, providing broader opportunities for students to complete tasks in authentic contexts. In open assessments, students may use resources such as notes, course materials and other tools, including AI, while completing the task.
The automatic text below will appear in the Assessment Details section of published course profiles based on the classification:
- Secure Assessment: This assessment is designed to protect academic integrity and assess your understanding of course concepts. You must only use the approved materials and tools for this assessment task. Use of prohibited materials in this task may be investigated and treated as academic misconduct. See assessment instructions for more details.
- Open assessment: In this assessment you can use resources such as your notes, course materials, and other tools, including AI, while completing the task.
From Semester 2 2026, this automated text replaced the requirement to manually add standard AI-use text to course profiles.
From semester 1, 2027 the assessment procedure requires secure assessment tasks to be more significant components of assessment in every course with:
- secure assessment worth at least 60% of the course grade; or
- a secure assessment item that accounts for 30% or more of the overall course grade and acts as a hurdle requirement to pass.
As we transition to this level of security, courses that can not reach this level of security are asked to have an approved exception from ADA's and include the statement in the Assessment Additional information section for the relevant Course Instance:
This course has an approved exception to UQ’s usual secure assessment requirements. Its assessment has been reviewed in the context of the course and the broader program. For more information, see clause 18 of the Assessment Procedure
The assessment condition AI required has also been introduced and should be used when students are required to use AI and are assessed on their use of AI (e.g., tool selection, prompting, verification, and integration) within an assessment task. More details are available on the Course Profile assessment attributes page.
Student responsibilities and use of AI
Students are responsible for the quality and integrity of all work they submit.
Where AI or Machine Translation use is permitted, expectations for whether and how this use should be acknowledged must be clearly specified in the assessment instructions and aligned with disciplinary practice.
Minimal guardrails for AI use
Shared trust, shared boundaries
Staff and students can use UQ authorised AI tools to explore ideas, practise, develop work and support learning and teaching within the conditions of the activity. Responsible AI use depends on shared trust and clear boundaries. The minimum guardrails below set out what must be protected, what people remain responsible for, and the boundaries that apply to everyone. They provide a common starting point for using AI with confidence. Within these boundaries, the broader principles of responsible AI help staff and students decide whether, when and how AI is appropriate in their particular context.
Students and educators at UQ work within non-negotiable boundaries that apply across all four domains of the responsible AI use framework (PDF, 87.4 KB). These guardrails are also provided in a print format (PDF, 795.2 KB).
Guardrail | Educators and students should ... | Why this matters | UQ responsibility anchor |
|---|---|---|---|
Protect personal and student data | NOT upload identifiable student, educator or other personal data to public AI tools.
| Personal data must only be shared where UQ privacy, confidentiality and information-control obligations can be met. | |
Protect secure assessment materials | NOT upload unpublished, secure or sensitive assessment materials to public AI tools. | Assessment content must be protected to preserve academic integrity, fairness and institutional confidence. | |
Keep humans accountable for outcomes | NOT use AI as the decision-maker. | AI may support judgement, but accountability for assessment and student outcomes must remain with authorised UQ educators.
| |
Respect Indigenous knowledge governance | NOT generate, reproduce or represent Indigenous knowledge, culture, customs or traditions without appropriate authority, consent and governance. | Indigenous knowledge must be treated with respect, cultural authority and recognition of cultural and intellectual sovereignty. | |
Respect copyright, IP, confidentiality and licensing | NOT use AI in ways that breach copyright, licensing, privacy, confidentiality, UQ intellectual property, student intellectual property, or cultural and intellectual sovereignty obligations. | AI use does not remove existing legal, ethical or institutional responsibilities. | Intellectual Property Policy; Privacy Management Procedure; UQ Reconciliation Action Plan; Library collection notice |
Staff and students may use UQ’s authorised AI systems (e.g. Copilot) with permitted materials and in conjunction with human judgment, to:
- Explore ideas and possibilities by brainstorming, generating examples, considering alternative perspectives and developing initial plans or outlines.
- Build understanding by asking questions, seeking plain-language explanations, connecting concepts and exploring unfamiliar topics.
- Support practice and revision by generating practice questions, flashcards, worked examples or opportunities to apply and test knowledge.
- Summarise and organise information from notes, readings, documents, meetings or other authorised sources to support understanding and planning.
- Analyse and compare information to identify themes, similarities, differences and areas requiring further investigation.
- Develop and refine communication by creating initial drafts or improving the clarity, structure, grammar, tone or accessibility of existing content.
- Create draft learning and teaching materials, including examples, activities, presentation outlines, discussion prompts and assessment resources, for educator review and refinement.
- Explore the implications of AI for a discipline or profession, including how AI-generated outputs should be evaluated, verified, disclosed and used responsibly.
- Critique AI-generated outputs by checking accuracy, evidence, bias, missing context, invented information and suitability for the intended audience.
- Support inclusive participation, including through translation, alternative explanations and different ways of organising or presenting information, while checking that meaning and disciplinary accuracy are preserved.
- Complete assessment using AI in accordance with the instructions provided by course teams.
Acknowledging and referencing AI
Expectations for disclosing, acknowledging or citing AI use are context-dependent. Assessment instructions should require disclosure only where it is relevant to the task and disciplinary practice and should state clearly what students need to provide.
Guides and templates for acknowledging and citing AI
- UQ Library Guide: Acknowledge use of ChatGPT or other AI
- UQ Library Guide: Cite or reference use of ChatGPT or other AI
- Example: Acknowledgement slides (PPTX, 246.3 KB) developed by Dr Caitlin D'Gluyas, School of Social Science
- Example: Assignment coversheet information and coversheet template based on this guidance.
- Example: BIOM3020 AI Acknowledgment Coversheet developed by Dr Lisa Akison, School of Biomedical Science
- Claude.AI acknowledgement artifact you can adapt for your course (requires a free Claude.AI license to customise)
As UQ educators, we must recognise that our students encounter different AI acknowledgment expectations across their coursework, reflecting the diverse disciplinary practices within our institution. We need to explicitly identify and model the expectations for students to acknowledge AI use in each of our courses.
If an assessment task allows the use of AI, you need to set expectations with your students to either:
- reference AI using citation practices, like APA or MLA style guides
- acknowledge AI using a form, template or cover sheet
- both describe how you use AI (acknowledge) and cite your AI use (reference).
In general, if AI is the primary source then it should be cited, otherwise citing the primary source material AI uses is generally preferable (APA, 2024). If AI is used to edit, summarise or help plan, we should expect an acknowledgement. More complex uses of AI may be best described in a methods section or appendix with details of the process, tools and prompts used.
Your decision should be informed by your disciplinary norms. Most style guides provide a means to cite AI tools, with the guidance being updated as AI tools change.
AI detection
AI detection tools, including Turnitin, are flawed and unreliable. They cannot definitively determine whether text, ideas, or structures were generated by AI. These concerns have been raised through research and the popular press (Bridgeman et al., 2025). At UQ we had the Turnitin AI detection tool available for staff to use, this tool was disabled in mid 2025.
As AI becomes more prevalent across society and work, there is an increasing expectation that graduates will be comfortable in using this technology and be able to leverage those capabilities to support their work, where appropriate.
There is increasing evidence that AI detection tools, including the Turnitin AI Writing Indicator, generate false positives and disproportionately flag certain cohorts of students. Using these tools can also undermine our messaging that students will need to engage responsibly with AI to prepare for their graduate roles.
It is important that UQ staff not upload student work to external AI detection tools, as this risks, again, undermining our approach, but also raises data privacy risks.
To minimise reliance on AI detection, UQ is adopting more secure assessments where responsible AI use is evident. This approach fosters integrity and ensures fairness while adapting to the evolving role of AI in education.
Talking with students about AI, assessment and integrity
UQ students want clear guidance on how to use GenAI properly in their courses (Students’ Guidance for Academics, 2024).
60% of UQ students were discouraged from using AI because they were worried about breaking their University’s rules (Student Perspectives on AI survey).
We need to talk with our students about how to use AI in our assessment to help them learn to use AI ethically and to give them the confidence needed to safely to explore and learn to use it effectively.
Clarify course expectations
- When are students permitted to use AI in their assessment?
- Highlight the information specified in the course profile and explain why these decisions were made.
- How should students acknowledge and/or Cite AI use?
- Provide clear expectations, model these practices and link to the relevant guides.
- Support available for using AI ethically and effectively.
- Identify the support available at UQ and further support, if provided, within your course.
- Identify the support available at UQ and further support, if provided, within your course.
Activities to spark discussion of AI, assessment and integrity
- Explore case studies of the use of AI beyond university – particularly where it relates to your assessment.
- Provide a space to discuss AI in your assessment in an online discussion board.
- Debate AI ethics in the context of your assessment.
- Facilitate group activities to identify and bust or confirm AI myths.
- Incorporate AI into class activities, practicing the use of these tools, and
- how to use AI ethically and effectively
- acknowledging or citing the use of AI.
As educators, it's important that we model academic integrity and talk to our students about why academic integrity is important and how to engage with academic integrity in their assessment.
Guide students to reflect on their AI use in Assessment
Encourage students to ask themselves these 3 questions for every assessment before submission:
- Is this assessment task my intellectual and academic work?
- Have I followed my course coordinator’s instructions on using and acknowledging AI?
- Am I submitting work that demonstrates my learning, skills and abilities?
If students answer "no" to any of these questions, they may risk breaching UQ’s AI rules and be using AI in ways that do not support their learning.
Shetty and Anantharaman (2025) provide an example of supporting first year students to develop AI literacy.
More resources to use with your students
Example guidance for students acknowledging AI use adapted from Dr Peter Worthy (PPTX).
The AI student hub provides:
Copyright and the use of AI tools in Teaching
Copyright and AI can be complex, many AI systems require user to provide rights to the content used. This page sets out a few use cases for how an AI tool might be used in a variety of teaching scenarios. This information only applies to the use of works with closed AI tools that have been authorised by UQ (Copilot). It does not apply to the use of AI tools that are free to use or have been used with a licence paid for by an individual.
There are two general uses here, one (a) which involves the use of third-party content with an AI tool, and one (b) that involves the use of personal/university owned works or works licenced with a Creative Commons licence with an AI tool.
The examples provided here consider our responsibilities with copyright, we also have many other responsibilities in teaching including: the quality of our teaching and resources, the way we make assessment judgments, and the software we use. Depending on the specific situation, a specific use might meet our copyright responsibilities, but this does not remove our responsibility to ensure the output is current, accurate and appropriate and we meet all the responsibilities of our roles.
a. Where third-party content is used
The use of third-party copyright works with AI tools without a licence or permission from the copyright holder is prohibited.
In most cases, content that comes from one of the subscriptions that the UQ Library has access to cannot be used with any AI tool.
The only exception to this is works that are licenced with a Creative Commons (CC) licence, as these licences do not prohibit the use of these works with AI tools.
The following is a list of acceptable uses of third-party content with an AI tool where UQ has an appropriate licence or permission:
Examples of uses:
- Creating a summary of third-party content for use in class/distribution to students
- Analysis of key themes, sentiment and findings in an article for use in class/distributing to students
- Synthesis of themes and information between articles – resulting in the creation of a new work for use in class/distribution to students
- Creation of revision questions for use in class/distribution to students
- Creation of class presentations, tutorials, lecture slides etc.
- Paraphrasing text from third-party content for use in class/distribution to students
- Creating infographics, flyers, posters, etc. based off third party content for use in class/distribution to students
- Transforming third-party written content into an audible format
- Creating a translation of an article, book chapter etc into a student’s native language to assist with understanding
- Lecturer organising group projects and tasks where third-party content is uploaded
- Producing a reference list for content distributed to students
- Checking a reference list for accuracy etc. before distributing to students
- Generating scenarios, examples, or case studies for distribution to students (for assessment or teaching)
- Creating teaching and learning activities based on theories, readings etc.
- Analysing data for patterns and trends for class use of distribution to students
- Creating data visualisations for use in class/distribution based on published data
- Curriculum analysis
- Converting third-party text content to video/podcast for use in class/distribution
b. Where third-party content is not used
The list below outlines some acceptable uses of works owned by UQ or works licenced with a Creative Commons licence with an AI tool.
Examples of uses
- Quiz test and exam question creation for use in class/distribution to students
- Checking the accuracy of and/or marking tests and exams
- Lecturer transcribing their own lecture into written text for distribution to students
- Creating a translation of a course content/assessment task into a student’s first language to assist with comprehension
- Testing the efficiency of equations, calculations or proposed solutions
- Producing a reference list for content distributed to students
- Checking a reference list for accuracy etc before distributing content to students
- Writing emails to students
- Generating AI content to critique in class against individual work
- Improving grammar and sentence structure
- Translating assessment feedback into native language to assist with understanding
- Translating individual written content into English
- Organising raw data/facts into spreadsheets and tables for use in class/distribution
- Creating data visualisations for use in class/distribution based on raw data
- Resource transformation for accessibility purposes
- Curriculum analysis
For seminars, workshops and info sessions related to UQ's Lead through Learning strategy (2025-2027).