LearnWise AI Pilot Project

Pilot Duration: September 2026 - May 2027
Platform Integration: Canvas LMS (in-course navigation and university-wide navigation)

As artificial intelligence reshapes the landscape of higher education, the University of Alberta is taking a deliberate, research-informed approach to digital transformation. The LearnWise AI Pilot Project is an initiative running through May 2027 designed to evaluate how grounded, privacy-focused virtual assistants can enhance teaching, streamline academic support, and improve student retention and engagement.

Rather than adopting unvetted commercial chatbots, this pilot integrates a university-dedicated virtual assistant directly within Canvas. By testing this technology over an extended period, the U of A aims to measure its operational impact, ensure alignment with our academic mission and establish long-term governance standards for emerging technologies.

Purpose

The primary goal of the pilot is to explore the benefits of artificial intelligence while maintaining institutional control over data, pedagogy and academic standards.

  • Data privacy and institutional security: LearnWise AI is hosted securely in European data centres and fully aligns with rigorous university security and privacy standards. Course materials, lecture notes, syllabi and student interactions are never used to train commercial AI models (such as ChatGPT or Claude) and are never shared externally.
  • Grounded, reliable responses: Unlike open-web AI tools susceptible to "hallucinations”, LearnWise AI operates on a "grounded" framework. In-course answers are drawn strictly from the verified materials uploaded by instructors, ensuring students receive accurate, course-specific guidance.
  • Proactive student support: By embedding assistance directly inside Canvas — where students already spend their academic time — course logistics, study aids and campus resources can be accessed at any hour of the day or night.

Information for instructors 

  • Automated logistics and administrative relief: LearnWise AI handles routine, repetitive student queries regarding assignment due dates, syllabus policies and course logistics, freeing up valuable office hours and email time for deeper academic mentoring.
  • Actionable course insights: The system aggregates student queries, providing faculty with an "Insights Dashboard" that highlights common misconceptions, recurring questions and areas where students may need additional support or clarified resources.
  • Pilot participation control: Faculty retain full authority over whether their course participates in the pilot. While in-course AI assistance is enabled by default, instructors can easily disable or customize the tool for any individual course via Course Settings → Navigation.
  • Restricted data access: To ensure AI support focuses strictly on course content rather than assessment, LearnWise is prohibited from accessing sensitive materials such as exams and quizzes.
  • Accelerated feedback in SpeedGrader: Instructors can utilize AI-assisted draft feedback aligned directly with assignment rubrics inside Canvas SpeedGrader. Every draft requires human-in-the-loop review, editing and approval — keeping instructors in full control of final grades and academic judgment while significantly reducing grading bottlenecks.

Information for students

  • 24/7 in-course academic support: Students receive instant help with concept clarification, exam preparation and study prompt generation based directly on their instructor's uploaded syllabus and course content.
  • Centralized campus navigation: Through the floating global assistant, students can navigate non-academic campus services anytime, getting quick answers for IT support, library databases, financial aid deadlines and administrative processes.
  • Clear academic integrity guardrails: The platform is engineered to support learning, not replace it. Built-in guardrails prevent the AI from drafting essays, completing assignments or answering quiz and exam questions.

Ongoing evaluation and community feedback

A successful pilot relies on the lived experiences of our teaching staff, support personnel and student body. Throughout the duration of the pilot until May 2027, the project team will actively collect qualitative and quantitative feedback through multiple channels:

  • Faculty and staff surveys: Regular check-ins to measure time savings, feedback quality in SpeedGrader and overall satisfaction with course-level controls.
  • Student Experience Assessments: Evaluations focusing on ease of use, helpfulness of study assistance and impact on course navigation.
  • Focus Groups & Advisory Sessions: Dedicated sessions with faculty committees, student government representatives and instructional design teams to discuss emerging needs and potential platform refinements.
  • Usage & Resolution Analytics: Reviewing anonymized, aggregate metrics to assess deflection of routine support tickets, usage patterns during peak midterm/final periods and overall platform reliability.

Get involved and learn more

Whether you are an instructor looking to test AI-assisted grading, a staff member managing department questions, or a student seeking to make the most of your Canvas tools, resources are available to help you navigate LearnWise AI:

  • Faculty resources: Visit the Faculty AI Resource Page for step-by-step guides on testing the tutor, managing course navigation settings and optimizing rubric feedback in SpeedGrader.
  • Student resources: Check out the Student Guide to LearnWise for best practices on using AI responsibly to aid your studying and campus navigation.
  • Configuring the pilot tools: Visit IST’s knowledgebase article. 
  • Feedback and inquiries: Have thoughts or questions about the pilot? Reach out to our project implementation team at lmsproject@ualberta.ca.