Academic Integrity and Artificial Intelligence
At UniSA, we believe that we need to thoughtfully consider all aspects of teaching and learning and how the principles of academic integrity can be applied in relation to not only generative AI but all technology.
Generative artificial intelligence (AI) is a rapidly evolving technology (e.g., ChatGPT, Sustain AI, Dall-E). This technology is purported to create original content on which the system has been pre-trained based on prompts that users provide. The content can be refined until the user deems it fit for purpose.
The explosive interest in these tools is leading academic staff to question the validity and reliability of their assessments and has raised concerns over academic integrity. Currently, we lack sophisticated detection tools. Turnitin has announced software that can help determine if AI has been used to compose text-based assessments. They anticipate the tool will be available in April 2023. Open AI, the creators of ChatGPT, has also released a text detection tool.
For more information about Academic Integrity, please refer to our newly updated Academic Integrity Policy (AB-69) and Academic Integrity Procedure (AB-69 P1).
What we know about AI and what it can do
The quality of the output of Generative AI tools varies dependent on the size and quality of the data set that they were fed, the amount of pre-training conducted, the prompts provided and their programmed functionality.
The prompts users can provide to Generative AI tools can include not only a question to be answered or a creative premise to expand upon, but the prompter can also provide information on the context. These contextual elements can include demographic information of the intended audience and the author, the purpose of the piece, emotional quality or even asking it to adopt the style of a specific author.
Recent media attention associated with such tools has focused on assessment integrity and the broader impact on the higher education sector.
UniSA both recognises the proliferation of such tools as an opportunity and acknowledges the risks and weaknesses. Such tools are and will increasingly be common in our and our students’ personal and professional lives; therefore, we need to develop informed judgement about when and how to use such tools.
The Australian Academic Integrity Network (AAIN) has made available AAIN Generative Artificial Intelligence (AI) Guidelines. The document provides guidance on the appropriate use of generative AI in higher education aligned with the Higher Education Standards Framework (1.4.4).
How will this impact academic integrity?
The impact of these technologies on academic integrity will depend on the context in which they are used. There could be instances where their use is necessary and important to developing the skills our students need for their future academic and professional careers.
There will also be instances where writing, image and multimedia creation and other endeavours are either the focus of the learning or the process of creating these works helps students to develop critical thinking, writing or creation skills as well as subject matter mastery.
We need to ensure students have multiple opportunities to demonstrate the attainment of course objectives across a program. Therefore, applying good assessment principles, as defined in the Authentic Assessment definition - valid, reliable, and fair, provides feedback and promotes academic integrity.
Actions
We recommend reviewing our assessments and consider the following actions:
- Find out more about the technology what they can and can't do
- Engage in conversations with your students about ethics and academic integrity and using generative AI
- Consider if and how industry is using these tools and the skills students will need to support their career goals
- Communicate expectations on the use of AI in assessments
- Re-consider your current assessments
- Remind students that the Higher Education Standards Framework (1.4.4) contains expectations for students
- Remind students of their obligations as per UniSA's Academic Integrity Policy (AB-69)
Summary
- Explore and understand the basics of how generative AI can be used – read this explainer blog
- Engage your students in conversations about using artificial intelligence
- Communicate expectations on using AI in assessments
- Make explicit to students the definition of Academic misconduct as per UniSA's Academic Integrity Policy (AB-69)
- Re-assess assessments to ensure they are consistent with the learning outcomes being assessed
- Use Turnitin and other available software (new software for AI coming in April 2023)
Resources and links
- Teaching Students to Write with AI: The SPACE Framework
This article by Glenn Kleiman, a Senior Advisor at the Stanford Graduate School of Education not only suggests a framework to help students understand how to write with AI, the article also explores the limitations of AI tools such as ChatGPT. Kleiman provides an overview of the technical limitations as well as warning of the biases inherent in the publicly available data and documents general-purpose AI tools use as sources.
- ChatGPT for Educators
Some hints and tips for educators from Dr. Philippa Hardman, Cambridge University Scholar and consultant on how educators can use AIs for a range of educational and administrative activities.
Also consider Dr Hardman’s guide to Writing Learning Objectives in a Post-AI World.
- Chat GPT & Education
A series of slides that describe and demonstrate what ChatGPT can do for educators from K-12 up through higher education. Also includes overview information of generative AIs, how to talk with students about using AIs and provides resources for further reading. Written in January 2023 by Dr Torrey Trust of UMass Amherst.
- The ChatGPT Prompt Book
Dr Alan D Thompson provides a series of slides exploring how to write effective basic and advanced prompts for ChatGPT for a range of circumstances from grocery lists to coding languages whilst providing an overview of the uses and limitations of the tool. Dr. Thompson’s newsletter and website explore a range of AI tools.
- AI in Education: Guidance for Policy Makers
This big picture view of the use of AI in Education from UNESCO published in 2021 offers some interesting insights into the use of AI in Teaching and Learning as well as AI’s use in Educational Administration. It also considers if and how AI can support UN Sustainable Development Goal https://sdgs.un.org/goals 4: Ensure inclusive and equitable quality education and promote lifelong learning opportunities for all.
- Practical Responses to ChatGPT and Other Generative AI
This information page provides a range of information resources about Generative AIs, their use in Teaching and Learning and some practical strategies for design and pedagogical approaches to address their use in formative and summative assessments and learning tasks. Shared under a Creative Commons 4.0 licence.
- Teaching and Learning in the Age of AI: what are the opportunities?
A presentation by Dr Vitomir Kovanovic, Senior Lecturer, UniSA Education Futures as part of the TIU Breakfast Series on
Thursday 2 March 2023.
- Best Practices for Using AI When Writing Scientific Manuscripts
From the American Chemical Society. A look at the strengths and weaknesses of ChatGPT and a set of recommendations of best practices for scientists when using such tools at any stage of research, particularly at the manuscript writing stage.
- Using Generative AI to Develop Your Research Questions (video, 03:45)
This short video provides an example of how to use AI to develop starting research questions.
- AI and Scholarly Publishing: A View from Three Experts
The article covers a panel interview from late October 2022 but still has timely information and an overview of the use of AI in scholarly publishing. The information shared covers a range of ways in which AIs can support scholarly writing and publishing and provides descriptions of tools and their uses and limitations. The panel consists of Helen King, Head of Transformation at SAGE Publishers, Dr Lucy Lu Wang, Assistant Professor at the University of Washington Information School and visiting researcher at the Allen Institute for AI and Dr Paul Growth, Professor of Algorithmic Data Science at the University of Amsterdam, and scientific director of the UvA Data Science Centre.