With broad access to generative AI (GenAI), much of the focus has been directed towards answering the questions “what does GenAI mean for students?” and “how will professors reconcile the use of GenAI in the classroom?”. However, these questions leave out a crucial member of our teaching and learning community: the GA/TAs. This section aims to provide an introduction to GenAI, including how it may impact GA/TAs in their roles. You can also find answers to commonly asked questions through our frequently asked questions page.

What is Generative Artificial Intelligence?

Generative AI is a type of artificial intelligence that is capable of generating text, code, images, and video. GenAI includes a number of different types of AI models, including Large Language Models (LLMs), which are powerful and useful tools that can generate text in response to written (and sometimes video or image) prompts. Though these programs may seem to create answers to our most important questions by “thinking about them”, they are not capable of true thought. Instead, these models have been developed by examining massive amounts of data, which act as a foundation of knowledge for the GenAI. They then use algorithms to construct sentences in response to these prompts one word at a time. Words are selected by identifying the most likely word to appear next in each sentence based on their training data, any fine tuning of the model, any supplementary resources (such as documents or websites), and the context provided to them by the user. Most are now also able to access the internet to check for knowledge not contained in their training data. By doing this, GenAI can generate reasonable responses to many questions, and many specialized AI systems can check their answers on the fly against other sources of knowledge. However, in many cases, especially with free versions of publicly available AI, they are unable to answer questions with adequate detail, nuance, and clarity or may hallucinate responses, so it is important to carefully consider the output of AI tools and not just accept it.

How Might GA/TAs Encounter GenAI?

GA/TAs hold a great deal of responsibility in many courses through leading labs and tutorials, grading assessments, hosting office hours, and a whole lot more. As a result, it may be helpful to consider the following ways that you may encounter GenAI in your role:

• GA/TAs may encounter GenAI outputs in courses with unsupervised written assessments, where students can input the assignment instructions as a prompt and receive a complete assignment, aligned to the grading criteria.
• The availability of wearable AI, including more affordable versions of AI systems, means that courses may experience the impact of widely available GenAI. As such, in-person assessments (e.g., oral assessment) can be compromised, even when handwritten.
• GenAI is now highly capable of multimodal, mathematical, symbolic, image and video reasoning, and may be able to complete any undergraduate, and even many graduate level assignments, to a very high level. On top of this, the public release of agentic AI models means that students can direct teams of AI agents to autonomously complete many tasks on their behalf.

The wide availability of GenAI can be overwhelming in your role as a GA/TA, erode trust and cause massive disruption to how we understand and facilitate teaching, learning, and the creation of knowledge. To detect potential GenAI responses in assessments, keep an eye out for hallucinated references or parts of prompts and AI responses left in submitted text. It’s also important to fully understand any allowable uses of AI and what those might look like, and to avoid letting AI mistrust make it seem like all student submissions are AI generated when they may not be.

Can GA/TAs Use GenAI to Complete Required Work?  

GenAI is a powerful tool but should not be used as a way to outsource GA/TA responsibilities as it undermines the very position GA/TAs are compensated for. Grading is one of the most important tasks you will undertake as a GA/TA, and while it is often time-consuming, it’s critical that grading and provision of feedback are done with integrity and to a high ethical standard.

There are already many teaching technologies (for example, from textbook publishers or specialized tools for specific fields) that use AI to grade student work and/or provide feedback, and these tools will become increasingly common. GA/TAs are likely to encounter these systems in the capacity of a ‘human in the loop’ checking the accuracy of grades and feedback, dealing with student questions and so on. This is a specifically trained and constrained use of AI. This usage is very different than submitting student work to a public GenAI tool like ChatGPT, Gemini, or Claude, which should be avoided for several reasons. Firstly, these public GenAI tools may not provide accurate or reliable responses for your particular context and may provide biased or incomplete answers. Secondly, doing this may lead to the student’s intellectual property being used to train future models without their consent which is problematic.

Legitimate Uses of GenAI in GA/TA Work

It is possible that GenAI may be legitimately and intentionally designed into the teaching, learning, and assessment within the classes you support, and in those cases, you may be required or encouraged to use AI for certain tasks. If that is the case, your instructor should provide very clear and explicit instructions about what their expectations of you are.

Below are a few examples of how you might use AI to support learning. Though before using GenAI to assist with any GA/TA related duties, this should be discussed with the course instructor to ensure transparency, integrity, and appropriate usage.

Generating alternative explanations: You may find in grading students’ work that there is a consistent misconception you are seeing. AI can be very helpful in creating alternative explanations or even a model answer to share with students.

Checking tone, clarity, and grammar of student comments: AI can be helpful in checking the tone, clarity, and grammar of messages (e.g., feedback, emails) you plan to send to students. If you are struggling to write a piece of feedback in a supportive way, AI can offer suggested phrasing.

Lesson planning: If you are tasked with running tutorials or review sessions, it can help you generate ideas for activities to support learning, or different and creative ways to explain particularly tricky concepts. You might be able to create a Google NotebookLM project for the class that acts as a learning support or review document (generating podcasts, review questions, flash cards, slides, summaries, mind maps and more). You also have access to Microsoft Copilot chat in Office 365, which can be helpful in working with Excel spreadsheets (e.g. if you are tracking grades externally), improving written documents, enhancing PowerPoint slides, finding emails, drafting email responses and so on. Copilot is the only institutionally supported enterprise AI tool available at present, and it keeps any data you share with it inside the University’s environment, so is safer to use than many other public alternatives.

In all cases, it is worth having a conversation with the course instructor before using AI in ways that can impact outcomes for students, or which may involve sharing the instructors, or others’ copyright materials with AI tools.

Adopting an Ethical Approach to GenAI

GenAI is a powerful tool despite its limitations, and those limitations are also rapidly changing, so GA/TAs need to be adaptable and aware of those changes. Given the ubiquity of access to GenAI, it is important to adopt an ethical approach to its utilization in higher education that aligns to the University’s institutional values for the responsible use of AI.

1. Maintain a high degree of integrity when performing GA/TA duties.

It may be tempting to use GenAI to directly grade students’ assignments or provide feedback – you may even think that AI is able to do a better job than you in some cases. However, using general purpose public GenAI is problematic for the reasons outlined above, and requires some skill to be able to reliably use this way.

Grading is an opportunity for GA/ TAs to provide actionable feedback on assessments to undergraduate students, often based on their own experiences and knowledge. This allows the student to develop an understanding of the quality of their work and where they will be able to improve for the next assessment. But this benefit is not limited to the students; providing actionable and high-quality feedback allows you to develop skills such as effective communication, providing guidance, and identifying correct and incorrect information. These skills are both highly sought after and transferable to many careers and can only be developed through actively engaging in grading.

2. Communicate with your instructor about GenAI

As of March 2024, the University of Windsor created resources for course instructors which guide the use of GenAI in the classroom. These guidelines give instructors the option of implementing policies regarding GenAI for their own courses. Therefore, each course will require GA/TAs to understand the policies set by the instructor they are working with. Bylaws 54 and 55 require course instructors to include a statement regarding acceptable and unacceptable uses of GenAI in their course. As such, all GA/ TAs should review and understand the relevant policies and procedures regarding GenAI by consulting the course syllabus and discussing expectations with the instructor prior to commencement of GATA duties.

3. Model ethical use of AI to students

If, in consultation with your instructor you decide to use GenAI, it is critical to model ethical and transparent usage:

• Respect the copyright of both course instructors and students. Understand that public, and especially free GenAI systems may utilize the information from any materials that are input to the system for commercial purposes, such as training or enhancing AI models. For this reason, some course instructors may not wish for their personal materials to be included as prompts for the development of materials for GA/TA duties. Communicate with your course instructors regarding the use of their materials and provide this same information to students. It may still be possible to find and use alternative materials such as openly licensed (e.g. Creative Commons) content as a base for your prompts.
• If you have an agreement with the course instructor to use materials created by GenAI, ensure that those materials are submitted prior to use in GA/TA duties for review and approval by your instructor.
• Be mindful of the biases that can come from the data that GenAI have been trained on, the ethical position of the parent corporations of these systems, and review all materials for accuracy while also adhering to the policies outlined by the course instructor.