Student and instructor access to Generative AI (GenAI) raises new questions around how to effectively teach, support, and evaluate student learning. Given this, it is important that teaching assistants (TAs) have a clear understanding of institutional guidelines, course policies, and best pedagogical practices in light of this emerging technology. This resource offers guidance for TAs on how to best navigate the challenges and opportunities presented by GenAI, in order to work with a course instructor (CI) to create a consistent experience for students working with the technology in a course.
This document relies heavily on the guidance given in the Generative Artificial Intelligence in the Classroom: FAQs. Please note that the FAQs are subject to change, and we recommend reviewing them regularly.
About Generative AI
What is GenAI and how does this shape your role as a TA?
As a TA, understanding the capabilities and limitations of GenAI (particularly Large Language Models, LLMs), and engaging in a conversation with a CI about their integration in a course is crucial for effectively guiding students and using these tools into the educational environment. They are trained to predict the next word in a sequence, allowing them to generate coherent text. Their key features include the ability to maintain long-range context, meaning they can pay attention to earlier parts of the text to ensure coherence over longer passages. Additionally, their exposure to billions of pages of text during training enables them to generate diverse and contextually (e.g., prompt-dependent) relevant responses, based on carefully crafted prompts. Given these considerations, how GenAI is integrated into the course (i.e., content, engagements, and/or assessments) will depend on the CI’s course design approach.
One significant drawback is that GenAI may generate plausible-sounding but incorrect information, so it is vital to verify facts using reliable sources. Furthermore, these models may reflect biases present in their training data, which requires critical evaluation of their outputs. In addition, generative AI tools sometimes struggle in stating facts or quotations accurately. This means that these tools may generate claims that sound real, but are clearly wrong.
Educators are welcome and encouraged to test Microsoft Copilot, ChatGPT, Gemini, Claude, Perplexity and other tools, use of which are currently free. You can also test other AI tools to assess their capability, for instance to see how they respond to the assignments used in your courses, the way in which they improve the readability and grammar of a paragraph, or the way they provide answers to typical questions students may have about course concepts. Experimentation is also useful to assess the limitations of the tools. However, confidential information should never be entered into unprotected AI tools. Content entered into ChatGPT, Gemini, or other public, unprotected tools may become part of the tool’s dataset.
TA Tip: Explore the various resources on the U of T’s Artificial Intelligence site, especially those pertaining to Tools & Training and Resources for instructors and teaching assistants.
As a TA, it is always good practice to use institutionally approved tools. Currently, Microsoft Copilot is the recommended generative AI tool to use at U of T. When a user signs in using University credentials (UTORid and password), Microsoft Copilot conforms to U of T’s privacy and security standards (i.e., does not share any data with Microsoft or any other company). It is also free to use for U of T students, staff, and faculty. Microsoft Copilot uses OpenAI’s GPT-4 model and performs comparably to ChatGPT. For more information about Copilot, refer to CTSI’s Copilot Tool Guide.
TA Tip: Given that generative AI technologies are advancing each day, continue to refer to the OVPIUE’s FAQ page for new information and guidelines in the U of T context, as they arise.
Communicating with your Course Instructor about GenAI
What questions should I ask my course instructor (CI) during the pre-course meeting?
The pre-course meeting is a crucial time for you to connect with the teaching team about the teaching materials, course policies, and the people involved in the delivery of the course. Because GenAI is new technology and instructor vary in how they allow students to use GenAI tools, it is crucial to understand the CI’s expectations. To develop a clear line of communication and have a comprehensive understanding of GenAI expectations for the course in which you are teaching, you can ask the course instructor a range of potential questions:
- AI usage guidelines: What are the specific policies around AI usage in the course? How may I best communicate these guidelines to students? What specific tasks will GenAI tools be used for in the course? What best practices should I follow to ensure the effective and ethical use of GenAI in my teaching?
- AI-aware assessments: How have assessments been modified to account for generative AI capabilities? Are students permitted to use AI tools to complete assessments? If so, how should GenAI generated work be cited?
- Grading as a TA: If assessments have been modified to incorporate GenAI tools or to address GenAI capabilities, is that reflected in grading rubrics? Can GenAI tools be leveraged in the grading process without submitting student work to any third-party software system for grading?
- Academic integrity: How is the CI planning to educate students about the responsible use of GenAI? How can I communicate with students about how to uphold academic integrity when working on assessments? Given that U of T does not support the use of detection tools for generative AI, what is the protocol for identifying possible cases of using GenAI as an unauthorized aid?
- AI tool familiarity: Do I need to familiarize myself with any specific AI tools that may be relevant to the course content, assignments, or student learning? Where can I get supports and/or training on using the institutionally supported GenAI tools? What technical support is available if I encounter issues with the GenAI tools?
- AI literacy: Discuss potential strategies for helping students develop AI literacy skills in the course, including how to critically evaluate AI-generated content and use AI tools responsibly.
- Ongoing communication: Consider setting up a system for regular check-ins with the course instructor to discuss AI-related developments or challenges throughout the term. Ensure that these check-ins are allotted in the DDAH Form.
TA Tip: If your course instructor currently has no policy on generative AI use, consider sharing the University’s sample statements for instructors to include in course syllabi and course assignments. By including this messaging in the course syllabus and assessments, you and your students will better know whether AI technology is, or is not, allowed. These statements may be used for both graduate and undergraduate level courses.
Content Development and Grading Considerations
As a TA, can I use generative AI to create pedagogical content?
TAs and CIs are permitted to use generative AI tools to develop or update content for course activities. If you want to utilize AI to generate teaching and learning content, be transparent and consult with the CI about a range of considerations for doing so. For instance, TAs may wish to use generative AI to create tutorial case studies or discussion questions to support students’ active learning. The TA could then chat about the generated case studies and questions with the CI. TAs who make substantive use of these tools are encouraged to acknowledge their sources and discuss this with their students. By doing this, TAs are modeling for students how to responsibly use GenAI tools.
If you use GenAI tools to create content as a TA, keep in mind that you may not own or hold copyright over the generated works. In addition, avoid inputting confidential information or intellectual property that you do not have rights to use, including student work or questions without permission. Content entered may become part of the tool’s dataset and could resurface in response to other prompts (except when using tools like the protected version of Microsoft Copilot).
As a TA, can I use generative AI to help grade assessments?
Please note that the instructor is ultimately responsible for ensuring the grade accurately reflects the quality of the student’s work, regardless of the tool used. As a TA, your course instructor can clarify expectations about how they would like grading to be carried out. Before using any AI assistance for grading and/or feedback, you must discuss this with your course instructor.
Among the factors your instructor will be considering includes guidance about protecting students’ intellectual property, the use of third-party tools, and transparency. The University asks that TAs and CIs do not submit student work to any third-party software system for grading, or any other purpose, unless the software is approved by the University. This is because a completed assignment, or any student work, is the student’s intellectual property (IP), and should be treated with care. The University currently has several licensed software tools available for facilitating grading, such as SpeedGrader and Crowdmark. These systems safeguard the student’s IP, and in the future may include AI-powered grading assistance.
By manually grading assessments as a TA, you develop crucial skills in analysis, feedback, and student-support. This process allows you to identify areas where students need clarification, and it helps to maintain a personal connection with students, leading to a more supportive learning environment. These experiences and skills are invaluable for your professional development.
As a TA, how can I detect the unauthorized use of generative AI in an assessment?
The University does not support the use of AI-detection software programs on student work. None of these software programs have been found to be sufficiently reliable, and they are known to incorrectly flag instances of AI use in human-written content. Some of the AI-detection software programs assess if a piece of writing was generated by AI simply on its level of sophistication. Sharing your students’ work with these software programs without their permission also raises a range of privacy, intellectual property, and ethical concerns.
As a TA, instead of using GenAI detection tools, consider using traditional methods for detection of potential academic misconduct, as outlined by your divisional academic integrity office. Educate students on academic integrity implications for using ChatGPT or other generative AI tool on a marked assessment.
TA Tip: Before marking assessments, be sure to establish expectations with the course instructor, fellow teaching assistants, and students around appropriate generative AI use.
Communicating with Students about GenAI
How should I guide students on how to use GenAI tools as learning aids rather than unauthorized aids?
As a TA, it is important to engage students in a conversation about what it means to use GenAI tools as learning aids, in contrast to being an unauthorized aid. The university allows for students to use GenAI tools as learning aids, for example, to summarize information or test their understanding of a topic. These tools should be used in a manner similar to consulting library books, online sources, peers, or a tutor. Such uses are generally acceptable even if an instructor has stated that AI tools are not otherwise permitted in the course. These uses typically do not need to be cited or disclosed. In contrast, the university expects students to complete assignments on their own, without any outside assistance, unless otherwise specified in assignment instructions.
If your role as a TA involves communicating with students in tutorials, labs, or office hours, consider dedicating ample, ongoing time to discussing and modeling appropriate GenAI use. If your CI is permitting students to use GenAI tools on their assignments, discuss what tools and what use is acceptable: Can students use it for critiquing their work? For editing? For creating an outline? For summarizing sources? For searching the literature (e.g., using Semantic Scholar)? It is important to continuously emphasize to students that generative AI models can sometimes produce hallucinations, disseminate incorrect information, or generate biased outputs, highlighting the importance of continuously reflecting on and reviewing the generated content.
How should I communicate with students about copyright considerations when using GenAI to create study and learning aids?
As a TA, let students know that they may use AI tools as learning aids (e.g., quizzes, flashcards, summaries) as long as they comply with copyright and intellectual property guidelines and the instructions provided by the course instructor. In general, course materials should not be uploaded to GenAI systems, because some AI tools use uploaded documents or prompts to train their models. Students may violate copyright laws if they submit course materials, such as lecture slides and syllabi, to third-party tools without owning the IP or copyright. However, students can use the institutionally-approved version of Microsoft Copilot to generate study and learning aids, because of the data protection offered through this tool. Given this, consider giving a demonstration in class on how to use Microsoft Copilot as an effective learning aid.
There remains significant legal uncertainty concerning the use of generative AI tools in regard to copyright. This is an evolving area, and our understanding will develop as new policies, regulations, and case law becomes settled. If you have further questions about copyright, please view the U of T Libraries webpage, Generative AI tools and Copyright Considerations for the latest information.
TA Tip: As a TA, consider sharing relevant campus resources for students to support their learning skills development while using generative AI, thereby encouraging academic integrity:
- UTSG:Through the Centre for Learning Strategy Support (CLSS), students may book a learning strategy appointment or attend a workshop on effectively preparing for assessments and exams. At a Writing Centre, students may book writing support consultations.
- UTM:Through the Student Resource Hub at the Robert Gillespie Academic Skills Centre, students may book a learning strategy or writing support appointment.
- UTSC:Through the Centre for Teaching and Learning, students may book a writing support appointment. Through the Academic Advising and Career Centre, students may book a study-skills peer coaching appointment or attend a related workshop.