AI and the Writing Process: Overview
All Resources in this Series
What is Generative AI, such as ChatGPT?
Generative AI (GenAI), such as ChatGPT, uses algorithms and large-language models (LLMs) to predict and generate text that looks and sounds like a human has written it. The technology draws from a huge repository of existing text on the internet. It does not create new information or knowledge, but it summarizes, synthesizes, and re-words existing information.
In ChatGPT, for example, the user writes a prompt (an instruction or a question) and the AI returns a generated response very quickly. If the user gives the AI subsequent prompts within the same chat, then ChatGPT amends and adjusts its answers based on the whole conversation. This means the user can follow up to ask for more detail, for expansion on a particular element, or for a response in a different style.
This resource and associated resources will focus on ChatGPT for its examples and guidelines. We have used ChatGPT-3.5 as the free and accessible version. Other software exists, both free and with paid access, but it is beyond the scope of this resource to outline the ever-changing landscape of GenAI tools.
It is your responsibility to be the critical thinker in your interactions with GenAI.
What can ChatGPT do?
It can produce content in a huge range of styles, from poetry to the style of a well-known person or character. It can respond to questions and prompts with general content drawn from internet sources. It generates text in a variety of languages, and it can generate programming code.
What can ChatGPT not do?
It cannot identify what is true. Its goal is to sound like a human, so it makes things up, from facts to sources of information. GenAI cannot think or write. By this, we mean that, by synthesizing from a huge library of texts, it can generate true-sounding text without any kind of critical understanding about whether that information is true or biased. It will not tell you if it is repeating someone else’s ideas, and if you ask for sources, it will often manufacture or make up fake references.
Again, users need to think critically about all content that GenAI generates.
Should I use it?
It depends. Two words should guide your use of GenAI: productive and ethical.
Productive means that your use of GenAI is helping you learn. It is supporting you in developing your thinking, processing, language, and writing skills. Productive use of GenAI means using it as a tool to help you as you are completing work and not as a substitute for you doing the work. Much depends on your learning stage and learning goals, since too much reliance on GenAI, even as a tool, can be unproductive for learning.
Ethical means that you have permission to use GenAI and you are transparent about how and where you have used GenAI in your work through documentation and citation. You explain when you used GenAI as a tool, what it contributed to your work, and what you did with the information it provided. Citation of GenAI use is one part of this.
Cautions and Risks
As with any technology or tools you use, you should be aware of ethical concerns and other cautions associated with GenAI to make an informed decision about whether to use these platforms.
- ChatGPT and similar GenAI draw on available information from across the web, but the content outputted is not attributed to the people who created or wrote the source materials. Users should be aware of this ethical and potentially legal issue, and they should conduct research to confirm and source ideas and information found through ChatGPT.
- In addition to the above, there are legal and copyright concerns associated with using GenAI in particular ways. The Library’s copyright guidance on GenAI encourages caution with GenAI use.
- If you use or enter any of your own work into a platform like ChatGPT, you should be aware of and comfortable with the privacy permissions you’re giving for the platform to collect, use, or learn from your data and inputs. Read the terms and conditions for any platforms you use.
- GenAI can replicate social biases or miss critical balancing information. Users should use critical thinking to counter biases. Users should also conduct research to check the information and perspectives provided by the AI.
- ChatGPT and similar GenAI defaults to generating text in Standard American English (SAE), which reduces linguistic diversity and stylistic variety. If you use ChatGPT as a model for your own writing or speaking or to give you examples to build from, consider how your voice, your languages, and your choices make your writing individual and unique.
- Open AI, the company that developed ChatGPT, has been criticized for unethical and exploitative labour practices to clean its datasets of toxic content and hate-information.1
How can I use it?
The resources in this series outline ways that GenAI can support your drafting and writing processes.
Please note that, if you’re not permitted to use GenAI for a course or assignment or you’d prefer not to, there are many other learning support options. In addition to asking your instructors and TAs for help, other services, including the Writing and Communication Centre (WCC), the Library, AccessAbility Services, and the Student Success Office (SSO), can give you guidance for planning, researching, and drafting your assignments. The WCC can also help you get started and give you feedback as you draft.
There are many kinds of GenAI and AI tools. ChatGPT can seem like a great tool for many tasks, but it’s not always the best tool. Check with the support services listed above for recommendations on the best tools to use for academic tasks like research, citation management, data visualization, graphics creation, etc.
Before Using GenAI for Course Assignments
Academic and professional integrity are critical elements of student and professional conduct. Always check that your use of GenAI is allowed.
Check assignment guidelines and be responsible about documenting and citing GenAI use.
For course assignments, some course instructors will allow and encourage students to use GenAI within a set of parameters they provide in the syllabus or in assignment instructions. Some instructors prohibit use of GenAI for course assignments.
Without the explicit permission or instruction of your instructor, you should never submit work produced by ChatGPT or other AI. Doing so is an academic offense. From University of Waterloo guidelines: “Using ChatGPT (or similar tools that generate text, code, or visual images) for content generation and submitting it as one’s own original work is a violation of the University of Waterloo’s Policy 71 (Student Discipline).”
To demonstrate academic integrity, always follow your instructor’s guidelines for using, documenting, and citing GenAI contributions to your research, drafting, and design processes. Refer to the linked resource on documenting and citing GenAI use. If you have any questions about where and how you are allowed to use AI, always ask your instructor or TA.
Next Steps: Linked Resources
The following linked resources provide more information on using GenAI as a student at the University of Waterloo. These resources follow the general writing process for drafting a written assignment and they outline specific examples for how to incorporate GenAI as a tool for each stage.
1 Perrigo, B. (2023, January 18). “Exclusive: OpenAI Used Kenyan Workers on Less Than $2 Per Hour to Make ChatGPT Less Toxic.” Time. https://time.com/6247678/openai-chatgpt-kenya-workers/