Generative AI Tools
Tool Comparison
This page helps you choose generative AI tools you can use safely and securely in your teaching. It details which tools the University of Toronto supports, their differentiating features, available resources and supports, and universal considerations for using any AI tool for teaching and learning purposes. The guidance follows the principles in the University’s AI Task Force Teaching & Learning Working Group Report (June 2025): student-centred design, transparency, equitable access, and respect for academic freedom.
This guide focuses on tools that belong to the University’s AI Kitchen, a secure environment for AI adoption. Tools in the AI Kitchen are vetted for privacy and security and covered by U of T’s data protection standards: your prompts and your students’ data are never used to train AI models, and you can use institutional data up to Level 3, such as course materials and student assessments.
The AI Kitchen currently includes a suite of tools that have use cases for teaching and learning:
- Cogniti is the preferred tool for teaching and learning. It is built for education: embedded in Quercus, free for every student, highly customizable, and can be grounded in your course content.
- Microsoft Copilot Chat is the easiest place to start. It is a general-purpose tool available to everyone at U of T and is the lowest-barrier way for students to build basic AI literacy.
- Licensed Tools in the AI Kitchen (such as Claude for Education, ChatGPT Edu and M365 Copilot) may suit your own course preparation and other projects where advanced features can add value. However, some of these tools require a paid subscription that is not covered by the University.
Looking for AI-powered tools for literature searching and research? See the overview of AI-powered tools on the University of Toronto Libraries’ AI Academic Search Engines resource.
To compare the supported teaching and learning tools, use the table and decision guide below.
Tool Comparison
| Dimension | Cogniti | Microsoft Copilot Chat | Licensed Tools |
|---|---|---|---|
| Licensing and Cost | Free. No cost to instructors or students. | Free. Included in U of T’s Microsoft 365 license. | Claude for Education is free. ChatGPT Edu and M365 Copilot (an additional license that extends the functionality of the free Copilot Chat) are not centrally funded. |
| Accounts and Access | Accessible to all instructors and students via Quercus. No new account required. | Accessible to all instructors and students via U of T’s Copilot web portal. No new account required. | Both instructors and students need licenses. Custom chatbots cannot be shared with unlicensed users. |
| Teaching Use Cases | Virtual tutors and study assistants grounded in course content, role-play and simulation, formative feedback, and student-built chatbots. Optimized for integration into course activities. | Brainstorming, drafting and revising, summarizing concepts, and generating discussion questions. | Advanced capabilities and custom chatbots for course preparation, administrative, and professional work. |
| Customization and Educational Alignment | Full control of system prompt, course materials, persona, and pedagogy (for example, Socratic questioning). Built-in system prompt templates and settings that limit non-course-related usage. | General purpose chatbot with no customization. | General purpose with customization for chatbots. No built-in educational alignment features. |
| Learning Analytics and Student Feedback | A dashboard provides an aggregated summary of student conversation themes. Individual, deidentified conversations can be reviewed and exported. Students can provide feedback on chatbot responses for instructor review. | None. | None. |
| Quercus Integration | Yes. Embedded directly in Quercus pages. | No. | No. |
| Privacy and Data Security | Vetted. Meets U of T data protection and privacy standards; prompts and student data are not used for model training; institutional data can be used up to Level 3. | ||
| Safety Guardrails | Vendor-supplied safety guardrails, plus an institutional safety layer maintained by CTSI: directs students in distress to U of T mental health resources; refers students to Navi for general inquiries about U of T supports and resources. | Vendor-supplied safety guardrails. | Vendor-supplied safety guardrails. |
Resources and Supports
| Dimension | Cogniti | Microsoft Copilot Chat | Licensed Tools |
|---|---|---|---|
| Getting Started | Refer to CTSI’s Cogniti Tool Guide for instructions for how to get started with Cogniti, including enabling it in Quercus courses and requesting the ability to create agents. | Refer to CTSI’s Microsoft Copilot Tool Guide. | Refer to the AI Kitchen website for up-to-date getting started information. |
| Supports Available |
CTSI hosts regular Cogniti workshops that cover getting started, chatbot creation, and customization: see our GenAI Workshops page. Contact q.help@utoronto.ca or book a consultation with CTSI for support with Cogniti. |
Contact q.help@utoronto.ca or book a consultation with CTSI for support with teaching and learning applications of Copilot Chat. Contact the AI Kitchen support team for technical support. |
Contact q.help@utoronto.ca or book a consultation with CTSI for support with teaching and learning applications for paid subscription tools in the AI Kitchen. Contact the AI Kitchen support team for technical support. |
Considering a Tool Outside the AI Kitchen?
Any AI tool the University has not vetted or licensed, such as personal ChatGPT, Claude or Gemini accounts, open-source models, discipline-specific tools, or custom chatbots on outside platforms, fall outside the AI Kitchen. Such tools are subject to their own terms of use and often use customer data for model training and other commercial purposes. Several considerations apply before using such tools for teaching and learning:
- Students must be able to opt out. You cannot require students to accept a third party’s terms or hand over personal data or intellectual property. If the tool is part of a required activity or assessment, you must provide a viable alternative for students who decline.
- A Privacy Impact Assessment (PIA) is required wherever personally identifying information is collected. Under U of T’s FIPPA obligations, any unvetted tool that collects personal information (e.g., student names, email addresses, assessment submissions) needs a PIA to be conducted. This process typically takes four to eight weeks from the time all required information has been received. Refer to Information Security’s PIA Page for more information.
- Support. You are responsible for all technical support, training, and troubleshooting for yourself, your TAs, and your students. The AI Kitchen and CTSI do not provide end-user support for non-supported tools.
- Tool Changes. Tools may change their terms of service, pricing, or functionality at any time. A tool you have built your course around may become unavailable or fundamentally change mid-term.
What to Communicate to Students
Use of any non-vetted tool should be explicitly described on the course syllabus:
- State clearly what tool is being used and for what purpose.
- Indicate whether students are required to create an account, and advise them not to reuse their UTORid password.
- Provide links to the tool’s privacy policy and terms of service.
- If the tool is used for a required activity or assessment, provide a viable alternative for students who do not consent.
- Remind students not to include personal information in their prompts.
- Address any known accessibility barriers.
For more information, see CTSI’s guidance on tools beyond Quercus.
Learn More
Institutional guidance
- AI Task Force Teaching & Learning Working Group Report
- AI Kitchen and the AI Kitchen Service Catalog
- Privacy Impact Assessments
- U of T Information Security – Use AI Intelligently
- Generative AI in the Classroom FAQ (OVPTL)
Teaching and learning support
- CTSI – Teaching and Generative AI at U of T
- Tool guides: Microsoft Copilot · Cogniti · ChatGPT Edu
- GenAI Literacy Course Modules
- Using AI tools for learning at U of T – UofT Student Life
- U of T Library – Generative AI and Copyright · AI Academic Search Engines
- U of T Mental Health Resources
This page is a living resource and will be updated as institutional policies, tools, and best practices evolve.
Artificial Intelligence Virtual Tutor Initiative
Cogniti is a chatbot powered by generative AI that is tailored to specific course materials. It provides students with responsive, on-demand support for teaching and learning, offering immediate assistance when students need help understanding course content.
CTSI supports instructors in designing and implementing course-specific virtual tutors. Instructors collaborate with support staff and peers throughout the design, deployment, and evaluation phases to build effective virtual tutoring practices.
To learn more about how to use Cogniti in your course, visit CTSI’s Cogniti Tool Guide.
If you have any questions, contact: ctsi.teaching@utoronto.ca
Meet Cogniti: Your Course-Specific AI Platform
As part of the current phase of the Virtual Tutor Initiative, the University of Toronto is piloting the use of Cogniti: similar to ChatGPT or Microsoft Copilot, but specifically tailored to course content and integrated within Quercus. Cogniti is designed to support learning by providing helpful responses to course content-related questions.
The University has thoroughly vetted the platform for data privacy and security. Please note that the accuracy and relevance of the answers will vary depending on the prompt used and the performance of the AI Virtual Tutor, which may evolve over the term as the technology advances and we refine the settings of the model.
The virtual tutor ”has the potential to foster equity in the learning environment if used by novice learners, allowing them to learn at their own pace."
Emily Ho, Assistant Professor, Occupational Science & Occupational Therapy
Resources and Support for Instructors
Downloadable Resources
- AI Virtual Tutors – Effective Prompting Strategies: Offers evidence-based prompting techniques and sample questions to help students get the most out of AI virtual tutors, supporting personalized learning, critical thinking, and deeper understanding of course material.
Web page | PDF | Word - AI Virtual Tutors – Developing an Effective System Prompt: Offers strategies to translate your teaching goals into effective system prompts that scaffold student learning and maintain pedagogical integrity.
Web page | PDF | Word - Preparing Content for Custom AI Chatbots: Guidance on how to organize, clean, and format course materials for use with custom AI chatbots, ensuring accurate, relevant, and efficient AI responses.
Web page | PDF | Word
For more downloadable GenAI resources: CTSI Resources
Cogniti Workshops & Webinars
Dedicated sessions are available throughout 2026 for instructors interested in Cogniti:
- Introduction to Cogniti: U of T’s AI Tool to Support Student Learning
- Cogniti Onboarding: Setting Up Your AI Tool for Student Use
- Customizing Cogniti: Aligning for Different Learning Goals
How Generative AI Works
Generative AI refers to artificial intelligence that can generate new content, including text, images, and other media, based on predictive modeling. It uses machine learning algorithms, specifically neural networks, to process and learn from large datasets. Large language models (LLMs) are one class of generative AI, and they have the ability to generate human-like text.
Getting Starting with Generative AI
Are you interested in learning how generative AI works? A good introduction is provided by the Schwartz Reisman Institute for Technology and Society’s “What are LLMs and generative AI? A beginners’s guide to the technology turning heads”. You could also consider:
The responses you receive from a generative AI tool depend on the prompts you enter, and the further refining of these prompts, which takes practice. As Ethan Mollick said, “The lesson is that just using AI will teach you how to use AI.” (Working with AI: Two Paths to Prompting) To get started, we recommend you consider the following.
Be clear about what you want. Include detailed information in your prompt, including the desired format. “Write a paragraph about…” “Create an image containing…” Suggest a particular style (e.g., an academic essay or lab report) and include specific information you want to include (e.g., provide an outline or ordered steps for the prompt).
- If you’re not sure how to describe the style you want to emulate, the Wharton School at the University of Pennsylvania suggests pasting in a text example you like and asking the tool to describe the style Use that description in your own prompt for style.
- To learn more on prompt writing, see our Tool Guide under “How can I prompt with Copilot?.”
Be critical.
- Does the tool output meet your needs? What additional information is required? Generative AI is an interactive tool. Try different options and prompts to gauge the results, clear the prompt screen and try again. You will learn to refine your prompts and better discern what is most effective with practice.
- Generative AI tools can provide quick results that may appear correct, but looks can be deceiving. Tools such as ChatGPT can produce hallucinations or misleading and factually incorrect text. As with any text or visual analysis, we need to examine the results with a critical eye.
Before deciding whether and how to integrate Generative AI into your course, it is essential to have a solid understanding of its capabilities in supporting learning, as well as the challenges it presents. This section provides a basic overview of generative AI as it relates to teaching and learning.
Opportunities and Challenges of Generative AI in Classroom Learning
Tools that leverage generative artificial intelligence (GenAI) and large language models (LLMs) to generate new code or text (e.g., Copilot, ChatGPT, Claude, Gemini, etc.) are becoming increasingly available and are likely to have long-term impacts and on what and how we teach.
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. When intentionally leveraged for classroom instruction, generative AI technologies may also provide new possibilities for enhancing accessibility and engagement for students with varied learning needs.
When integrating GenAI into your courses, there are several opportunities to foster an inclusive learning environment:
- Encourage AI literacy and future readiness. As generative AI tools develop in capabilities, they will continue to shape what distinctive human skills are prioritized across disciplines and fields of work. By incorporating opportunities within courses to explore, use, and assess generative AI tools, students are better prepared to strategically engage with and critically reflect on these emerging technologies.
- Encourage metacognition, creativity, and critical thinking. By designing carefully constructed assessments and forms of active engagement, students have the opportunity to explore different viewpoints, self-reflect, and engage in analysis and knowledge synthesis.
- Address barriers to equity and accessibility. Generative AI tools can be leveraged to provide multiple options for motivating and engaging learners, for representing information, and for inviting learners to express and communicate. This can create learning experiences that are personalized to students’ diverse needs and abilities.
When integrating GenAI into your courses, there are several considerations to ensure equitable and responsible use, including:
- Availability. While many generative AI tools are currently freely available, their availability could change at any time. For any third-party software that is not approved 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 non-approved generative AI tools.
- Accuracy and bias. Text created by generative AI technology may be biased and may not be correct.
- Academic integrity. The University is discouraging the use of tools that claim to be able to detect AI generated text. See Generative Artificial Intelligence in the Classroom: FAQ’s for more information.
- Privacy and security. A version of Microsoft Copilot is currently available to the public (and U of T students), however 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 may not offer such privacy and data protections.
- Copyright and intellectual property. It is important to be mindful of what content is entered into generative AI platforms that do not have institutional data protections in place. Never input confidential information or intellectual property for which you do not have the rights or permissions. All content entered may become part of the tool’s dataset and may inadvertently resurface in response to other prompts. See the U of T Library’s Generative AI tools and Copyright Considerations for more information.
You can learn more about strategies to plan for and integrate generative AI into courses by visiting the Teach with Generative AI section of this website.
This work is licensed under a Creative Commons BY-NC-SA 4.0 International License.