Five principles for the effective ethical use of generative AI
Published by |
Last updated 06 November 2024
This guide introduces five principles to assist academics in transitioning their assessments in response to the new opportunities that GenAI opens up, as well as academic integrity challenges.
In February 2023, as it became clear that generative AI (GenAI) implied the need for a major, but as yet poorly understood, shift in teaching and assessment, TEQSA provided guidance on the appropriate use of generative AI aligned with the Higher Education Standards Framework 2021 (HESF):
The HESF states that providers assure the quality of teaching, learning, research, and research training, with respect to content and skills developed, assessment and determination of learning outcomes, and the mitigation of foreseeable risks to academic and research integrity. Providers should document their decisions and monitor their progress in addressing generative AI.
HESF
UTS took a position of promoting ‘Effective Ethical Engagement‘ and the following principles provide guidance for staff to translate this into practice. We’re suggesting principles rather than recipes, recognising that effective, ethical engagement can take many forms to suit the contexts in which we teach. You and your colleagues will undoubtedly test many creative recipes to suit the needs and tastes of diverse students in diverse contexts.
This is clearly a significant transition for all educational institutions, and we’re all trying to make sense of the implications. To understand if GenAI will enhance or undermine learning, and what scaffolding students may need to use it effectively and ethically, you’ll need to explore the apps yourself, track how others in your field are adapting assessments, and test your assessment questions in GenAI apps prior to session starting.
The principles are framed in terms of the student experience we want to create, to help you determine what is appropriate and consistent for each course and subject, within institutional guidelines.
Student-centred principles
The following graphic details the five student-centred principles to guide usage of GenAI.
Students understand the significance of GenAI for society, careers, and studies.
Students understand legitimate use of GenAI in their studies.
Students are equipped to engage critically and ethically with GenAI.
Students experience GenAI’s strengths and limitations as aids to learning.
Students are assessed on what they need to know in an AI world.
Explore the principles
1. Students understand the significance of GenAI for society, careers, and studies
Why it’s important
GenAI is reshaping society, the workplace, and university study — for good and for ill. It will soon be present in every conceivable productivity tool and website. Students need to be introduced to this early on, so they appreciate how GenAI connects to their professions, and each course and subject.
What you can do
Update courses and subject content to include information about GenAI and how it is impacting disciplines/professions, including potential changes in the future.
Discuss GenAI openly with your students right at the start of your classes. Talk about how it is already being widely used in the workplace and society, how it is impacting your discipline and provide illustrations and invite their questions. You might model how you have used it yourself.
Explain to students that learning to use and critique GenAI is key to their education. Ask questions to help students understand the importance of developing the capability to think about, reason with and make judgements around the use of GenAI in preparation for future work.
Discuss the importance of integrity as an aspect of their education and a central part of their future professional lives. Address the issue of cheating.
2. Students understand legitimate use of GenAI in their studies
Why it’s important
Students need clear guidance on how to ethically and effectively engage with GenAI for a given assignment, without crossing the academic integrity line.
“Copying or reworking any material (e.g., text, images, music, video) from generative AI tools, and claiming this work as your own without declaring use of the relevant tool.
Using generative AI tools, unless permitted use is specified for that assessment.”
What you can do
As Course Directors and Subject Coordinators, your curriculum and assessment choices have direct implications for any academic integrity questions around academic misconduct that may arise.
Given the rich diversity in the University’s teaching practices and the tools we use, clearly one size cannot fit all. Clear guidance must be given on what is permitted for each assessment.
3. Students are equipped to engage critically and ethically with GenAI
Why it’s important
Students must be helped to develop an awareness of when and how to use different apps, effectively and ethically, to advance their learning.
GenAI can be impressive, but it’s far from perfect. Moreover, even if working well in a technical sense, it can undermine learning if students ‘outsource’ too much work to it.
While students must take responsibility for their learning, they expect UTS to help them develop this AI literacy, just as they expect any other educational technology to be used to advance learning.
GenAI functionality is appearing in mainstream productivity tools and search engines, as well as myriad web apps.
What you can do
Discuss how and when the use of GenAI will deepen learning versus undermining it. This is not necessarily straightforward, and UTS will be providing support to assist academics to think this through.
AI literacy for learning may well differ from AI literacy for a professional: a power tool for an expert is not necessarily an effective tool for a beginner. Is learning being undermined through premature or poor use of GenAI — or can learning be accelerated through effective use?
Resources
UTS
What does critical engagement with ChatGPT look like? Video excerpt, 4 mins: Insights from 4 UTS academics on the diversity of ability they’re seeing in their students’ use of ChatGPT (Simon Buckingham Shum, UTS)
GenAI literacy can be thought about as a mix of knowledge, skills and dispositions, e.g.
Knowledge: curriculum and/or wider understanding is needed to validate factual claims and calculations, spot significant omissions, etc. Do you expect your students to have the requisite knowledge, or must they research it?
Skills: e.g., writing skills and fluency with the tool (including prompting) are needed to improve on prose which may in fact exceed the student’s own ability. If AI-writing is better than many of your students’, can they be reasonably asked to improve on it?
Dispositions:e.g., agency, curiosity, resilience, authenticity — needed to resist the temptation to just run with what the AI serves up. How to give your students the confidence and commitment to stamp one’s own mark on the piece?
4. Students experience GenAI’s strengths and limitations as aids to learning
Why it’s important
Students must learn how to harness the strengths of an AI tool, and understand its limits.
A conversational agent such as ChatGPT has strengths and weaknesses. Both of these can be leveraged to so students learn to appreciate these, and in doing so, engage in deeper learning.
What you can do
Browse the resources to see some of the ways that UTS academics have integrated GenAI into their assessment tasks, plus other creative and evidence-based proposals for its use. Do any of these look like they could translate well into your context?
What does critical engagement with ChatGPT look like? Video excerpt, 4 mins: Insights from 4 UTS academics on the diversity of ability they’re seeing in their students’ use of ChatGPT (Simon Buckingham Shum, UTS)
5. Students are assessed on what they need to know in an AI world
Why it’s important
AI may change what students need to learn, how they learn, and how you assess them.
GenAI can now perform high level cognitive work formerly considered to be the province of human cognition.
You must decide if students should still master the same knowledge and skills you’ve always taught pre-AI, or if this requires a rethink of what students need to learn, and how they can best evidence this.
A 24/7 conversational aid may also open exciting new ways of learning that are humanly impossible to scale.
Understanding these strengths and weaknesses will take time and experimentation.
What you can do
Course Directors and Subject Coordinators should update curricula to include GenAI as a professional tool.
Assessments should reduce the risk of using GenAI for misconduct but also allow students to demonstrate they know how to use it in the context of their course.
Do students still need to demonstrate mastery without AI tools? Should authentic performance with AI tools become an explicit goal? Or are both now needed?
