FAQs - Chat GPT at the University of Waterloo

ChatGPT and similar technologies are artificial intelligence backed chatbots that can mimic human conversations and writing (O’Brian, 2023). ChatGPT is a Large Language Model that learns the statistical structure of language, such as patterns of word usage, to generate answers based on probability distribution over word sequences (Ramponi, 2022). As Chat GPT composes an answer, it determines the most likely word or sequence that should go next, based on the training that it has had to date. These tools can be used for a variety of tasks including drafting emails or blog posts, composing essays, and even generating, debugging, and documenting code. This technology is particularly powerful as it can mimic writing or coding styles relatively effectively, making it particularly flexible and widely applicable. Some see this technology as the next generation of word processing tools, like predictive text, grammar checkers, or a kind of “supercharged Clippy” (the old MS Word “help” tool) (Bruff, 2022). Indeed, Microsoft is now building ChatGPT into MS Teams, Word, and PowerPoint. Rival companies will be releasing their own generative AI chatbots imminently, and tools for nearly any purpose are already available or in development. 

ChatGPT provides a response based on a prompt. The current public release of ChatGPT uses information found freely on the internet (up to 2021) to generate responses to prompts. Chat GPT uses information found freely on the internet, but is not a search engine. It “learns” through access to information curated by researchers, developers, and ongoing input from its users—which may contain errors, misconceptions, or bias (Wu, 2022). Of course this field is developing rapidly, and at the time of publishing these guidelines, similar tools are being adopted by search engines like Google. Verification and triangulation of information is important if using ChatGPT, as it is unable to distinguish between important and unimportant errors (Ramponi, 2022). While it can – if asked – provide references for content found, they will often appear plausible but be incorrect (Montti, 2022).   

The responses generated by ChatGPT can be incorrect and may include bias (Wu, 2022). ChatGPT responses can contain bias inherent within the free, large database of internet it was trained on as well as the potential bias of those reviewing and selecting the text to include in the large database of text that ChatGPT uses to create its responses. One writing specialist tested ChatGPT’s response to a prompt and found that it reinforced rather than complicated a simplistic ideological stance – unlike Wikipedia or even Google searches (Mills, 2023). The specific phrasing of prompts impacts how the AI responds, including the degree of bias in the response (Squires, 2022). Other similar AI chatbots such as Google’s Sparrow and Galactica were unsuccessful in their releases for this very reason (Wu, 2022). 

The University of Waterloo expects that students will do their own learning, and not fabricate, falsify, or otherwise pass off others’ work as their own. There are, however, ways to use many tools ethically for learning. 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 Student Discipline Policy policy, but other uses of these tools may be explicitly allowed by the assignment guidelines, course outline, or otherwise specifically permitted by the instructor. Instructors should clearly communicate their expectations for use (or not) of generative AI tools in the course syllabus and assignment instructions.  

At this time, the University does not have standard boilerplate text for use in syllabi. Course outlines must include the current standard statements on academic integrity and Turnitin. Faculty are asked to explain in course outlines and assignment guidelines what is and is not permitted, just as we do for collaboration or use of secondary sources.  

There are many tools emerging to detect ChatGPT use. However, controlling the use of AI writing through surveillance or detection technology is not recommended; AI will continue to learn and if asked, will itself help to avoid the things its own architecture is using to detect it. The plagiarism software detector we use, TurnitIn, indicates plans for integrating AI detection into its software. AI detection software is not currently used or recommended at the University of Waterloo, nor is there currently a statement for using AI detection software in a course to monitor student AI use. 

The Centre for Teaching Excellence has collaborated with other academic support units to develop resources for instructors.

The Centre for Teaching Excellence is in the process of developing resources to support teaching and learning.

There are many possible uses of generative AI for assessment; ChatGPT is not fully accessible globally, and now has paid subscription models. If used in assessments or activities, plan alternatives. One option is to require students to use ChatGPT content provided by the instructor, instead of asking all students to sign up for this particular tool.  

AI, such as ChatGPT, generates text from other sources; it uses others’ ideas without proper attribution. If using permitted chatbots in teaching and learning, students should always check and verify information and cite the original sources instead of the AI. This kind of checking skill is, of course, a valuable outcome for university-level learning. Citation guides for generative AI text and/or original prompts is an emerging question for commonly used style guides; while some academic journals and associations are banning references to AI as “co-author,” the American Psychological Association (APA) and Modern Languages Association (MLA) are beginning to develop citation norms and guidelines. More information is available from the University of Waterloo Library, Writing and Communication Centre, and Centre for Teaching Excellence.  

AI tools like ChatGPT have functionality that may create efficiencies for planning instruction and producing content. This will in fact become inevitable as it is integrated into the MS Office suite. There are examples of ChatGPT being used to help ideate, plan, and draft instructional materials like topics, outlines, lesson plans, content, questions, and other instruction. The University of Sydney has prepared advice about how AI can be used meaningfully by teachers and students in 2023. 

The use of generative AI that is then submitted as though it is one’s own original work would violate academic integrity policy. In addition, please  take note of the Waterloo policy on editing of graduate theses.

CTE, Teaching Fellows, and others are gathering use cases and will host sessions to share information between instructors.  

Instructors are encouraged to contact Academic Support Units for support with AI: 

• For support with teaching and learning, contact  the Centre for Teaching Excellence (contact via Faculty-based liaisons) 

• For support with online courses, contact the Centre for Extended Learning (contact via the Agile Development Team) for online courses,  

The University of Waterloo is working on building a community knowledge base that will evolve with AI advancement. In the meantime, check the following recommended resources to monitor evolving information about ChatGPT in Higher Education: 

• Resources for exploring ChatGPT and higher education

• ChatGPT through an Education Lens 

• ChatGPT & Education

• AI Text Generators and Teaching Writing: Starting Points for Inquiry

• AI Readings | Academic & Collaborative Technologies (ACT)Academic & Collaborative Technologies (ACT) (utoronto.ca)