FAQs
Last updated: October 15, 2026
1. About LearnWise and How It Works
1.1 What does this pilot project entail?
Answer: The pilot involves three LearnWise AI features integrated with Canvas:
- The Virtual Campus Assistant tool for students that answers questions about campus resources and services. It has no access to course information.
- The Virtual Course Assistant that provides student study support grounded in published Canvas course materials. Instructors can also use it to draft learning activities and to review the topics students ask about most. This feature uses Retrieval-Augmented Generation (RAG), a technique in which a large language model (LLM) generates its responses using only relevant passages retrieved from the course content in Canvas.
- The Virtual Feedback Assistant, a trial feature for instructors that is intended to assist with formative assessments. Working inside SpeedGrader, the Feedback Assistant uses AI to draft rubric-aligned feedback for the instructor to review, edit and publish. Instructors must request the Virtual Feedback Assistant be activated by contacting eClass support (eclass@ualberta.ca) as it is not active by default.
1.2 What is the administrative and support work that LearnWise can help instructors with?
Answer: LearnWise is intended to assist with routine, repetitive administrative questions (such as due dates, syllabus policies and course logistics in large classes) so that instructor time can focus on high-value teaching. It also offers instructors insights on which topics students ask about most, and can support drafting learning activities. The Virtual Feedback Assistant can draft rubric-aligned comments on formative assessments that instructors review, edit and publish themselves. It is not intended to replace feedback on learning or core teaching tasks.
1.3 Will LearnWise answer student questions without the instructor reviewing and authorizing the responses?
Answer: Some of the tools in LearnWise require instructor review while others do not.
The Virtual Course Assistant answers student questions directly, without the instructor reviewing each response first. This tool removes administrative burden from instructors by answering routine, repetitive questions (such as due dates, syllabus policies, and course logistics in large classes), allowing instructors to focus their time on teaching. The Virtual Course Assistant also guides learning for students with responses grounded in the course's published content (rather than providing direct answers about specifics in course materials). Instructors do not review these responses, either.
Unlike public or open GenAI tools such as ChatGPT, the Virtual Course Assistant tool has been configured by the university with guardrails designed to guide learning rather than draft essays, complete assignments, or answer quiz and exam questions. It does not replace instructor feedback on learning or core teaching tasks. Instructors who are not comfortable with responses going to students without prior review can remove the Virtual Course Assistant from their course at any time through Course Settings → Navigation.
The Virtual Feedback Assistant works within SpeedGrader and can draft rubric-aligned feedback for the instructor on formative assessments. Those drafts require the instructor to review, edit and publish. Instructors must request the Virtual Feedback Assistant be activated as it is not active by default. Instructors can contact eClass support (eclass@ualberta.ca) to have it enabled.
1.4 Will the Virtual Course Assistant in LearnWise be prevented from responding to inappropriate questions about the instructor or the course (for example, “is the content in this course biased?” or “are the assessments fair?”)?
Answer: System prompts and guardrails are built into the tool, and the Virtual Course Assistant is configured to support learning based on the course's published materials – not pass judgments on the fairness or quality of course material.
1.5 Given the virtual assistant cannot read images, audio, or have access to class discussion, how can it provide reliable formative feedback to students?
Answer: The Virtual Course Assistant is not compatible with all courses which is why it is a tool instructors can turn off in their courses. The Virtual Course Assistant works from published text-based materials in Canvas. It cannot access embedded video or audio, image-only (scanned) PDFs, discussion posts, or unpublished or locked content. Where courses rely heavily on unscripted lectures, class discussions, or non-textual materials, the tool's grounding is limited to what is present as text in Canvas.
2. Participating & Opting Out
2.1 Why is the Virtual Course Assistant pilot set-up as ‘opt-out,’ rather than ‘opt-in’?
Answer: The Virtual Course Assistant is enabled by default to increase the likelihood that the pilot would include a broad range of courses, disciplines, and class sizes. This will give the university a more complete picture when deciding whether this type of tool warrants ongoing investment.
2.2 Why is the Virtual Feedback Assistant opt-in?
Answer: The Virtual Feedback Assistant performs grading-adjacent work and so is not active by default and requires instructors to opt-in. Instructors can contact eClass support (eclass@ualberta.ca) to have it enabled.
2.3 Is it possible to opt-out of LearnWise midway through a term?
Answer: Yes. Instructors can opt out of the Virtual Course Assistant in any course they teach, at any time, by removing LearnWise from the course navigation under Course Settings → Navigation. Additionally, instructors using the Virtual Feedback Assistant can stop using it at any time.
3. About the Pilot: Goals, Timing & Decision-Making
3.1 What are the goals of this pilot project?
Answer: The pilot is intended to help the university measure operational impact and establish whether this type of tool, grounded in course content in the LMS, warrants ongoing investment.
For instructors, LearnWise can reduce repetitive administrative email burdens (such as routine syllabus and due-date queries in large classes), and provide visibility into the questions students ask most often. For students, LearnWise aims to provide 24/7 course-grounded study support, with configurations designed to guide learning rather than complete assignments for students. It also gives students a secure, institutionally supported AI pathway.
3.2 Why was the current project implemented when it was, right before the start of the Fall 2026 semester?
Answer: The LearnWise integration went live in Canvas as soon as it was finalized on August 18, 2026, so that the pilot could run throughout the full 2026-27 academic year. While we understand this did not provide much lead time ahead of the Fall term, launching before the Fall term means the pilot covers both Fall and Winter, giving the university a full year of instructor and student experience to draw on when assessing the pilot.
3.3 How did the decision to initiate this pilot project arise?
Answer: We know that instructors, researchers and students are already using open and free third-party AI tools like ChatGPT, and that those tools are not always well suited for a university learning and research environment - especially when it comes to security and control of information and protection of intellectual property. The pilot was designed to test whether a university-dedicated, privacy-focused assistant inside Canvas provides sufficient benefit to the learning environment to warrant ongoing investment.
3.4 Which committees or bodies discussed this pilot before it was implemented?
Answer: The initiative was discussed across advisory bodies, including the Digital Learning Environment Committee and the AI Steering Committee. The initiative was co-led with the eClass and Centre for Teaching and Learning teams. The GFC Committee on the Learning Environment and Student Affairs (CLESA) has received regular updates on Canvas and AI Literacy including AI tools, but not LearnWise specifically.
3.5 Were academic staff whose current job descriptions include teaching involved in the decision to start this pilot?
Answer: Yes. Advisory and governance bodies engaged in discussing the pilot include instructors, students and administrators.
4. Evaluating the Pilot
4.1 How will the success of the pilot be assessed?
Answer: The pilot is designed to provide the university with data to help inform if this type of secure, integrated tool warrants ongoing exploration and/or investment. The pilot will be deemed successful if it is determined to have produced enough valuable data to help make that assessment. Evaluation will draw on several sources including:
- anonymized, aggregate usage and resolution analytics to determine if a sufficient number of instructors and students are using it
- instructor and staff surveys
- feedback quality in SpeedGrader and satisfaction with course-level controls
- student experience assessments
- focus groups and advisory sessions with faculty committees, student representatives and instructional design teams
Surveys are being developed by CTL Scholars and interview details will be shared as soon as they have been confirmed. Information and details on the evaluation will be published to the university’s LearnWise pages. An anonymous feedback form is also open to the campus community. Feedback can be shared with the project implementation team any time by emailing ctllms@ualberta.ca.
4.2 What questions or metrics will guide this assessment, and will these questions be made available to the university community?
Answer: Yes, the university is committed to sharing components of the evaluation plan once finalized. The evaluation plan includes survey tools developed by CTL Scholars, along with the usage and resolution metrics described above. Information and details on the evaluation will be published to the university’s LearnWise pages.
4.3 Who, specifically, will make the decision about whether the pilot is successful or not, and will any instructors be involved in this process?
Answer: The decision will be joint between the Digital Learning Environment Committee and the AI Steering Committee.
4.4 When will this assessment take place?
Answer: Feedback is being collected throughout the pilot, and review and assessment will take place during and following the pilot. The Digital Learning Environment Committee will see a summary report of findings from usage metrics and the CTL Scholars’ survey and interview data findings to inform next steps as close to the end of the pilot as is feasible.
5. Intellectual Property (IP), Data Security & Jurisdiction
5.1 What access does LearnWise have to an instructor’s intellectual property?
Answer: LearnWise only accesses content that is already published and visible to students within the specific Canvas course site. It cannot access unpublished or locked content, grades, discussion posts, or quizzes and exams. Content from a previous offering of a course is only accessible if it has been copied into the current course and published.
LearnWise went through the Privacy Security Review process the university uses to ensure the security of data and IP for all digital systems, including Canvas. Under the university's agreement with LearnWise, course materials, lecture notes, syllabi, and student interactions are used only to provide the service to the university. They are not used to train AI models and are not shared with LearnWise's partners or other third parties for any other purpose. Data are held in the University of Alberta's own instance of LearnWise, hosted in European data centres.
5.2 If an instructor elects to enable the Virtual Course Assistant or the Virtual Feedback Assistant, what happens to the IP that these assistants ingest from Canvas to be able to function?
Answer: LearnWise only accesses content that is already published and visible to students within the specific Canvas course site. LearnWise went through the Privacy Security Review process the university uses to ensure the security of data and IP for all digital systems, including Canvas. Under the university's agreement with LearnWise, course materials, lecture notes, syllabi, and student interactions are used only to provide the service to the university. They are not used to train AI models and are not shared with LearnWise's partners or other third parties for any other purpose. Data is held in the University of Alberta's own instance of LearnWise, hosted in European data centres.
5.3 Who has access to course materials ingested by LearnWise?
Answer: Course materials accessed by LearnWise are available only to the instructor and students enrolled in that course. Students enrolled in the course receive answers grounded in its published course material. Students in other courses and users outside the university cannot access a course's materials through LearnWise.
Under the university's agreement, LearnWise may access course materials only as needed to provide and support the service, and may not use it for any other purpose. A limited number of university staff who administer the LearnWise integration may also have access for technical support. Evaluation of the pilot relies only on anonymized, aggregate metrics.
5.4 Would LearnWise use instructor-generated materials in the future for the same course taught by someone else other than the author?
Answer: No. LearnWise only accesses content that is already published and visible to students within the specific Canvas course site. Instructors maintain their intellectual property over their courses.
Copying of courses from one year to another is done through Canvas, not LearnWise. Only an instructor from a previous course is able to copy forward a course. If an instructor wants to share a course with a colleague, they should add them as an instructor in Canvas.
5.5 Who will have possession of the information generated by LearnWise in responding to students, given that this is new content created by AI after scraping a myriad of sources, which could include parts of assigned readings, as well as the instructor’s IP, like lesson plans, assignments, and rubrics? Who will control the dissemination/publication of this new content?
Answer: The instructor maintains their intellectual property. LearnWise does not disseminate information beyond responses to student questions grounded in visible course material. Student questions are visible to instructors.
5.6 LearnWise is an EU-based provider. Is LearnWise subject to EU laws/regulations over Canadian laws/regulations?
Answer: LearnWise is hosted in European data centres and is subject to EU data protection regulations (GDPR). The university’s Information and Privacy Office reviews all third-party software agreements, and the university's own obligations under Alberta privacy legislation continue to apply to U of A data regardless of where the vendor hosts it.
6. Virtual Feedback Assistant & Grading Integrity
6.1 Can the Virtual Feedback Assistant be used in high-stakes summative assessments?
Answer: The Virtual Feedback Assistant is a trial feature offered for formative feedback during the pilot. Formative feedback on lower-stakes work is where the tool may be helpful. While there is no technical restriction preventing its use on other assignments, it is not recommended for high-stakes summative assessment, where judging students' achievement of learning outcomes requires the professional judgment of instructors and TAs.
Instructors who use it should do so in line with the U of A's Framework for the Responsible Use of AI, in particular:
- Be transparent: let students know when and how the tool is used to give feedback on their work.
- Provide human oversight: review AI-generated feedback before relying on it. The tool is meant to support instructor judgment, not replace it, especially where decisions affect students.
6.2 Should TAs be allowed to use the Speed Grader/Virtual Feedback Assistant provided by LearnWise?
Answer: It is crucial that instructors who wish to experiment with the Feedback Assistant be involved in the work. Instructors can determine whether and how to involve TAs in this work. TAs should not be using these tools without the involvement of the course instructor.
6.3 If an instructor copies content from a previous section of their course to a new Canvas section, and some of the material that is part of this migration is no longer relevant to the current iteration of the course, could this cause issues when working with the feedback assistant?
Answer: The Virtual Feedback Assistant is grounded in the assignment rubric, so an outdated rubric would also affect its drafts. Instructors are advised to remove or unpublish outdated files and content, especially materials from previous terms. This practice is recommended to instructors regardless of LearnWise, as outdated course materials are still accessible to students. LearnWise does not have access to material not in the current course.
6.4 What process is in place to handle grade appeals when the Virtual Feedback Assistant, for various reasons, differs from the final assessment given by the Instructor (which may also involve input from TAs)?
Answer: Responsibility for grading rests solely with the human instructor. AI feedback is formative and is published to students only after the instructor has reviewed it. Any ensuing grade appeals would follow the normal process outlined in relevant policies and procedures.