Last Updated: 02 May, 2023 Views: 50
Please note: The use of Generative AI in academic contexts is an evolving field, and this advice may be subject to change.
- AI software tools can be useful, but they need to be used carefully and critically to make sure you are still producing a good standard of work.
- You need to be open and honest about the work you create with AI.
- It is very important to check the sources of content that AI generates, and do not assume they are correct (or even true).
What is generative AI?
Generative AI is a type of artificial intelligence that can generate new content (such as text, images, music, or video) based on a set of data it has access to.
Large Language Models (or LLMs), like ChatGPT, are "trained" to generate human-like written answers to written prompts using huge amounts of text. Text-to-image models, like Midjourney or DALL-E, use a large set of texts and images to generate "new" images based on written prompts.
AI for students
What can it be used for?
Within an academic context, generative AI is best used as assistive technology (do not use it to create content you are supposed to create yourself).
You can use generative AI tools to:
- Help you brainstorm ideas
- Improve your writing, for example spelling and sentence construction
- Write practice quizzes to help you revise for exams
- Checking the meaning of concepts, theories and methodologies
- Summarising information
- Generate decorative images for your PowerPoint
- Write background music or sounds for your presentation
And many other uses.
Can I use it?
You must check with your Subject Coordinator or Lecturer. If you are allowed to use AI you must be clear on exactly how you can use it and how to reference it correctly.
You must be careful not to contravene JCU’s Academic Integrity Policy - it is important to make sure your use of AI is openly acknowledged so that you don't pass off AI generated content as your own original work.
How do I use it?
Generative AI needs very specific questions (e.g. "tell me four reasons why diet can improve health outcomes for seniors")
Depending on the tool being used, the questions may need to be reworded to get better responses, or you can ask the program to improve on its responses ("can you change this to include a reference to Australian government standards?").
Generative AI is reliant on the prompts you use. While it does have some capacity to understand natural language, the better (and more specific) the prompt, the better the result will be. Learning how to use the tool well and structure good prompts will take time.
Is it free?
Some generative AI tools are free to use at present (but may be behind a paywall in the future), others have a free basic version and a paid subscription version, and others can only be accessed with a paid subscription.
It is important to consider the "other" ways you pay for access - like providing data. You may need to give your contact details and access to other personal data when you use a "free" tool, and conditions of use might allow companies to use or sell that data as they see fit. You should check the privacy policies and conditions of use before signing up to a service.
Generative AI does not necessarily cite its sources. You cannot tell where it gets its information from so you cannot check if it is accurate. This is the case with ChatGPT.
Even if it does provide sources, you should not trust them without checking them: It has been found to completely make things up- including references – this is referred to as “hallucinating”
It can only answer based on the data it has been trained on. It often does not have access to “live” websites or very recent information.
You are expected to submit your own original work for an assignment, and acknowledge when and where you use material that you did not create yourself.
You must be careful not to contravene JCU’s Academic Integrity Policy If you use AI content in your work without referencing it appropriately you will be penalized as you would for any other plagiarism. You cannot use AI to write your assignments. There is more information on the page Using Artificial Intelligence.
Bias and ethics
Always read the responses given by a generative AI tool critically.
Most software tools which are based on large sets of texts, like LLMs, have a tendency to reproduce biases from the source material.
As much of the content available for the models to "train" with is created by white Western men, the outputs generated by the tool can have a "world view" that is skewed towards this audience.
It is important to make sure you do not assume the content is culturally neutral because it was generated by a machine.
You should also be aware that the "ethics" of a computer program are based partly on the instructions and parameters set by the programmers, partly on the content of the data used to train the program, and partly on machine logic. This can be result in unpredictable and ethically "dubious" ethical standards being applied to the content the program generates.
Anything you enter into a generative AI becomes part of its training dataset.
There are no provisions for privacy regarding the material (e.g. personal photographs) you might use, or the questions/prompts you have asked. Content the software generates for you may also be used to "train" the software for future responses.
UNESCO ChatGPT and Artificial Intelligence in higher education: Quick start guide