What is Generative Artificial Intelligence and what application does it have for classroom instruction & learning?
Tools that leverage Generative Artificial Intelligence (AI) and Large Language Models to generate new code or text (e.g., ChatGPT) are becoming increasingly available and are likely to have long-term impacts and on what and how we teach. This resource is intended to offer some guidance around how to approach the use of such tools in your teaching practice.
You can learn more from FAQs in the University of Toronto document ChatGPT and Generative AI in the Classroom.
The University has created sample statements for instructors to include in course syllabi and course assignments to help shape the message to students about what AI technology is, or is not, allowed. Download the document here.
The University of Toronto’s Acacemic Integrity site provides some guidance on Using ChatGPT or other generative AI tool on a marked assessment.
University of Toronto Information Security has guidance on using artificial intelligence intelligently.
The availability and capabilities of generative AI technology are changing quickly. This page will be updated regularly. [Last update: March 28, 2024]
Higher education has faced similar disruptions with previous technology innovations, including calculators, Google search, and Wikipedia. While these innovations can be disruptive to our practices of teaching and assessment, incorporating them into our teaching practice is also an opportunity to prepare our learners to live and thrive in a changing world. The foundations of course design and authentic assessment can inform how we do this.
- While many generative AI tools are currently freely available, their availability can change at any time. As for any third-party software that is not supported by the University or your Division, there are several considerations related to privacy, security and student intellectual property that should be considered before asking your students to use generative AI tools.
- Text created by generative AI technology may be biased and may not be correct.
- Several tools that claim to be able to detect AI generated texts are now available. The University is discouraging their use for several reasons. See ChatGPT and Generative AI in the Classroom for more information.
New Resources from Inara Scott, J.D. , Gomo Family Professor and Associate Dean for Teaching and Learning, Oregon State University’s College of Business
- Yes, We Are In a ChatGPT Crisis
- Quick Responses To Chatgpt
- The “Less Content, More Application” Challenge
Other resources:
- Academic Experts Offer Advice on ChatGPT (Inside Higher Ed)
- Teaching and Learning with AI Apps (Taylor Institute for Teaching and Learning, University of Calgary)
- ChatGPT & Education (google doc from University of Massachusetts, Amherst)
- Adapting College Writing for the Age of Large Language Models such as ChatGPT: Some Next Steps for Educators (Critical AI)
- Embrace the Bot: Designing Writing Assignments in the Face of AI (Faculty Focus)
- The Office of the Vice-Provost, Innovations in Undergraduate Education, has issued a resource page on ChatGPT and generative AI in the classroom. This resource includes a FAQ section that addresses a variety of questions, including the use of AI-detectors such as GPTZero
- Connect with your Liaison Librarian as early as possible in your course. This assistance ensures you are taking a proactive, educational approach to designing assessments that make use of library resources – meet with them:
- For planning your assignments.
- For in-class presentations customized to your assignments addressing the information skills your students will need to be successful in your course.
- Integrate a divisional writing centre and its student supports into your course to ensure students are taught appropriate research writing and referencing/citation for your discipline.
- Regularly remind students of writing centre supports and the value of engaging with their support staff to build one’s confidence in research and writing.
Course and Assessment Design Considerations with AI in mind
Design for Learning Outcomes
- Alignment with Learning Outcomes – Consider whether or not the use of generative AI tools align with your course learning outcomes. Frequently share these course learning outcomes with your students.
- Leverage AI for Deeper Thinking - Consider how you might use generative AI tools to encourage problem solving, creativity, reflection, originality, and higher order thinking.
- Set Expectations around AI Tools – Be clear with your students about your expectations around their use of generative AI tools in the course and for each assignment. Discuss with your students what the tools do well and what they do not do well.
Webinar Recording: Content Creation Using AI: How AI Can Be Used to Build Courses and Learning Experiences
Dr. Stephen Murgatroyd hosts this engaging webinar sharing his experience on such topics as:
- Which AI resources are available that can help you quickly create quality content.
- What you should be cautious about and what we know are problems with AI-enabled content creation.
- How to leverage AI text-to-speech and speech-to-text for learning material creation.
- How to use the AI design features for a variety of material creation activities.
- How content curation is a major opportunity for AI-enabled supports, especially for students with exceptionalities.
Design Assessments keeping the use of AI Tools top of mind
- Scaffold or build assignments so that students are working towards a final product for submission. This effective approach benefits the students’ writing and learning and also creates authentic conditions that are more likely to deter use of generative AI.
- Check-in and communicate regularly with students during the assignment time period. This way, students will be less likely to seek out generative AI for (mis)information.
- Make assignments specific to your course experience. Base them on material covered in classes and tutorials (including class discussions and student presentations). Generative AI tools are also less able to address recent events or visual information. Also make clear that tests and exams will require mastery of work completed for assignments.
- Ask real questions that stem from current debates in your discipline, and let students know that you expect engaged critical thinking that is appropriate for the level of your students and your discipline. Encourage speculation based on evidence and reasoning, not just compilation of existing information or expression of unsupported personal opinion.
Consider a Variety of Assessments
- Presentations: in-person or virtual options: students (individually or in groups) submit an electronic ‘narrated presentation’ that can be assessed on the content and presentation techniques or present in class. Extend this assessment to include a reflection submission from all students with prompts specific to the presentations.
- Portfolios, logbooks, or assessment notebooks: students may submit a link to their electronic portfolios. Note that there is also an e-portfolio tool on Quercus that students can use to assemble their artifacts.
- Assessed seminars, group discussions and other similar activities (e.g., design frequent assessments to ask students to respond to a specific reading or discussion prompts).
Be Explicit about Expectations
- Provide rubrics and grading criteria to ensure students know what is expected of them in the assessment, also reducing inappropriate or misguided peer-peer communication.
Use a Universal Design for Learning Approach
- Anticipate the need for accommodation of students with disabilities completing assessments and be prepared to provide alternatives. For example, in-class writing assignments may not be appropriate for all students.
- Engage with colleagues in your discipline and across U of T to share approaches that integrate best practices in assignment and assessment design in the context of using or avoiding the use of generative AI.
Other Resources
- School of Graduate Studies’ Guidance on the Appropriate Use of Generative Artificial Intelligence in Graduate Theses
Enterprise Microsoft Copilot Artificial Intelligence Search (formerly Bing AI) Available to U of T Employees
On December 4, the University of Toronto will make available to all eligible faculty, librarians and staff a new enterprise version of an artificial intelligence-powered chatbot and search engine, Microsoft Copilot (formerly Bing AI). Microsoft Copilot responds to prompts with content and footnotes that link back to original sources on the internet. As with all generative AI tools, results may vary, and output may not be correct (sometimes known as “hallucinations”). Results should never be considered an authoritative source on a topic or an issue.
Please note that despite the name, this service is not the same as Microsoft 365 Copilot, and does not use or access Outlook emails, SharePoint files, Teams messages, etc. Chat data is not saved, nor is it used to train models or made available to Microsoft.
Students do not currently have access to the secure enterprise service. However, please note that some students are also employees; access to Copilot is connected to the primary license type assigned to individuals by their departments. Student-employees who are having difficulty accessing the enterprise Copilot experience should contact their departmental business officer directly.
Why is the University making this product available to employees?
While a version of Microsoft Copilot is currently available to the public (and U of T students), the public version does not have full privacy and data protections in place. U of T has access to the enterprise edition of Microsoft Copilot, which conforms to the University’s privacy and data protections, unlike the public version. Note that other publicly available generative AI chatbots like ChatGPT do not offer such privacy and data protections.
How does this product compare to ChatGPT or other generative AI chatbots?
While the Copilot AI search experience is based on the latest OpenAI models used by other applications like the free version of ChatGPT, because it is embedded within a proper search tool that includes references to real web-based sources, users are finding the experience to be better and more holistic with Copilot than other generative AI chatbots. In terms of security, using the enterprise version of Copilot better protects the privacy and security of end users. Remember that generative AI tools may provide information that is not correct (“hallucinations”), and it is up to each individual user to determine if the results are acceptable.
For more information and instructions on accessing the enterprise edition, please read and adhere to the guidelines for use: https://teaching.utoronto.ca/tool-guides/microsoft-copilot.
Please note the following requirements:
1. Use Microsoft’s Edge browser to access and use Microsoft Copilot. Currently, the service is not fully compatible with Chrome, Safari or Firefox.
2. Eligible faculty and staff must be logged into their U of T accounts to use the secure enterprise edition of Microsoft Copilot. If you aren’t logged in first, you will not be using the secure version. For detailed instructions on how to access and use the enterprise edition of the service, review the guidelines for use.
The enterprise edition of Microsoft Copilot has been evaluated by U of T’s Information Security team and is considered secure for use with level one and level two data. This service should not be used with private, sensitive, personal or confidential information, or material where the copyright holder has not granted permission.
If you have technical questions about Microsoft Copilot or other AI-related questions, contact q.help@utoronto.ca.
Sincerely,
Avi Hyman,
Director, Academic, Research & Collaborative Technologies,
Information Technology Services and the Centre for Teaching Support & Innovation