Generative AI is becoming increasingly common as it is integrated into browsers, existing websites and more. Like any tool, generative AI can improve productivity if used properly and ethically, but it also runs the risk of violating academic integrity policies if misused.
What is Generative AI?
Generative AI is a type of model trained on large quantities of source material that allow it to produce writing, images and more. Although there are a number of generative AI applications now available, one such type is based on Large Language Models (LLMs) capable of creating or editing various types of writing. These LLMs work by predicting the next word in a sentence, which it determines based on patterns in the texts that it was trained on. Due to how LLMs operate, they can produce fluent and natural-sounding language. However, it is important to recognize that these systems do not think, so they are prone to generating inaccurate claims and misinformation.
UofT’s Policy on Generative AI
Currently, the University of Toronto takes the stance that students are not allowed to use generative AI in a course unless the instructor explicitly permits it.
Instructors design their assessments to help students develop important skills for subsequent courses and for your future career, so unauthorized use of aids like generative AI can hinder your learning. Submitting your own, authentic work also provides TAs and instructors with a way to gauge your progress, which allows them to provide valuable feedback to help you improve.
If an instructor does permit the use of generative AI in their course, it is important to follow all of their guidelines if you decide to use it. In addition, you should not assume that an instructor will have the same policy on generative AI for all of the courses they teach.
Use of generative AI tools when not permitted is considered an academic offence and can result in an allegation of academic misconduct being raised by your instructor. If you are ever unsure about the generative AI policies in a course, always reach out to your instructor and ask first.
How do I know if I can use Generative AI in a course?
- Check the course syllabus and look for specific syllabus statements that indicate if you are permitted to use generative AI in any part of the course.
- Review instructions carefully for every assessment. Assignment instructions are often more detailed than the syllabus and rules regarding authorized aids often differ between assessments.
- If an instructor permits the use of generative AI, there may be additional rules that you are required to follow. For example, you may need to cite the portions that were generated by AI, or you may need to include the prompts and outputs that you used in an appendix.
- Paraphrasing apps and Grammarly’s AI writing partner service are derived from generative AI and some courses may not permit their use. It is always best to check in with the instructor before using these types of aids.
- Always ask your instructors if you are ever unclear about their policies on generative AI.
Cautions for Using Generative AI
As an emerging technology, the risks of generative AI are not fully known. Here are some helpful pointers to keep in mind as these applications become increasingly common.
Generative AI should support, but not supplant your learning.
Although generative AI can be a useful study tool, it is important that you improve your own skills and create your own understanding of the course material. This means that you should put in the time to do practice questions on your own or learn how to explain a concept in your own words.
Generative AI can be inaccurate.
While generative AI can create content that appears convincing, AI can produce misleading or false information. As such, it’s important to verify anything it generates, even if it seems legitimate. Look for primary sources that can support what it is claiming, and do not rely on it to do the research for you – remember that generative AI does not actually think, so it is not able to distinguish between factual or false information.
Do not input any private, confidential or identifying information into generative AI.
For many of these systems, the text you put in is retained and used to further train the AI. There is no easy way to have that information removed once it is in the system, so be careful about what you input – there is the risk that your information could be generated and leaked to others.
