Developing guidelines for generative AI in your subject

The UTS approach to the use of generative AI tools (for example, ChatGPT, DALL-E, Copilot, Midjourney) in learning and teaching is to encourage and support ethical, informed engagement. UTS courses and subjects will necessarily take a variety of approaches depending on their disciplines and contexts. By providing clear guidelines to your students you can support them to confidently engage with AI tools in an appropriate and responsible way.

In planning guidelines for your subject you will need to align with advice from your faculty and consider how AI intersects with your discipline area. We also recommend you consider the following:

• What are the learning outcomes for the subject? Appropriate AI use does not hinder students from learning what is intended by the subject; inappropriate AI use does.
• Should AI be allowed in assessment submissions, and if so how should students acknowledge its use?
• What are the emerging opportunities and impacts from AI in your discipline area? What skills will students need to critically engage with AI?

Talking about AI with students will give them the opportunity to ask questions and make suggestions. It may be useful to frame the conversation in relation to their learning and what they want from their studies, rather than focus only on academic integrity. If AI is likely to bring substantial change to how work is conducted in the discipline area, take the opportunity to talk this through and exchange ideas about these emerging changes, giving students assurance that their studies will remain relevant in this changing world.

Where assessment is concerned, it is best to provide guidelines in writing as well, making your expectations very clear and naming any specific forms of misuse that you want to call out. Students should not assume that use of generative AI for assessment is permitted unless specified by teaching staff. Students should also be aware that their assignments may be analysed for AI-generated text.

While it is clear that use of AI-produced content in an assignment without acknowledgement is a form of academic misconduct, what form the acknowledgement should take is less defined. We expect that some approaches will emerge and become standard practice, however at this stage there are a range of options and what is appropriate for a given subject, discipline and learning objective is up to you to define.

There is an ongoing discourse in higher education in relation to acknowledgement of AI contributions, some approaches to consider are:

• Citing their use through referencing. APA are in the process of deciding upon and formalising an approach to referencing of generative AI, in the meantime their interim advice is to treat the interaction with the AI as a form of personal communication, for example: OpenAI Chat GPT (personal communication, January 16, 2023), noting that personal communications do not appear in reference lists.
• A more detailed approach is to require students to provide all interactions with the generative AI including prompts and responses as an appendix. This approach offers the potential for the reader to understand exactly what the generative AI has contributed to the assessment.
• Another approach might be to ask students to critically reflect on their use of generative AI.

Resources and templates

• This Sharepoint site houses an Academic Integrity slide kit with talking points, and Canvas Commons templates for communicating Academic Integrity and GenAI use in your subject.
• Embed UTS Library’s new Generative AI Study Guide into your Canvas site – for help, contact your faculty librarian or contact the Library via chat or an enquiry