D2L Lumi AI Tools

• Faster feedback on assignments 
• More consistent comments across submissions 
• Additional suggestions that instructors can review, edit, and personalize 

• How to use Lumi Feedback
• Best practices for writing rubrics & instructions that work well with Lumi Feedback

Feedback is only available for submissions in the assignments tool right now (not discussions or quizzes/exams). When you click on a submitted assignment to grade, you will see a little icon pop up in the bottom left asking if you want to launch the Feedback tool.

• Create and refine course content 
• Generate assessment and discussion ideas 
• Save time on routine course development tasks 
• Enhance instructional design while keeping you in control of all final content 

• Create interactive content
• Generate materials from external files
• Generate Assignment ideas
• Generate Quiz Questions

You can find Pro throughout D2L in places where you see the sparkle icon indicating Pro features are available. Basically, think of Pro as something you can use to create new material or transform existing material. You can use Pro in the Content area to create new modules from a Word doc or PPT file, edit existing HTML content in a module, or generate module summaries based on content within the module. You can also use Pro to create/edit assignment prompts based on selected course materials, create/edit discussion topics based on selected materials, or create quiz questions in the question library based on selected learning content. In a nutshell, you can find it in Content, Assignments, Discussions, and Quizzes when you’re creating new or editing existing materials that you created in D2L!

(Note that you won’t see Pro when editing PDF or Word file uploads because those are uploaded materials rather than materials you created in D2L.)

One of the benefits of using Pro to upload Word docs or PPTs into your learning content is that Pro will adjust any existing accessibility issues you have in those file types. For example, if you are creating a module out of a PPT, before you generate that module from the PPT, Lumi will add alt text for any images within the PPT so the generated material will be accessible with a screenreader.

Note that you still may may many existing files within your D2L shell already that have inaccessible features (Word docs without heading styles or color contrast issues, PPTs without alt text for images, lecture videos without captions, etc.) If you’re wondering how best to fix those, here are some pointers: You can edit existing content that you created within D2L with Lumi Pro. But you can’t transform PDF or Words files with Lumi Pro because those are external file uploads, not material within D2L. The solution right now for adding accessible content from PDFs is to transform your PDF into a Word doc in Adobe, then choose to generate a new module from the Word doc you upload in D2L. Confused by this explanation? Join a workshop or schedule a Zoom with us, and we’ll walk you through it!

When multiple sections are combined into a single merged course shell, the AI assistant indexes and scopes its knowledge base uniformly to the unified container. It treats the combined materials, modules, and schedule within that single merged shell as one unified source for all enrolled students. If sections 2 and 3 are combined under section 1 and all learning material is set up in section 1 as normal, students in sections 2 and 3 will be able to use Lumi Tutor the same as students in section 1.

• Generate study materials
• Reinforce key concepts and course materials 
• Bolster critical thinking and self-directed learning 

Lumi Tutor is available for all students to access in D2L. If your concern is more about academic integrity than the tool itself, it’s worth noting that disabling Tutor doesn’t eliminate the underlying issue — students have access to general-purpose AI tools regardless. Many instructors find it more effective to set clear expectations about acceptable use rather than remove access altogether. 

Lumi Feedback and Lumi Pro are available for all faculty to access in D2L. However, you can choose not to launch any of the functionality either in your course (Pro) or the evaluation (Feedback) you give students. 

Lumi Chat is available for all faculty and students to use if they wish.

Any content generated by TCU faculty, staff, or students is owned by the TCU person who generated the content. Learnwise (the company that owns Lumi) and D2L do not own generated content, including activities. Nor is TCU’s use of the Lumi tools used to train AI models.

Tutor uses your course shell as its source of truth, rather than the internet at large. If you reference external content in your course shell, Tutor pulls out the content on the specific page you link to in your course, but it doesn’t crawl or follow links to other pages on that site. 

No, it doesn’t crawl or follow links to other pages on that site. So, if you want Tutor to reference one particular article or resource you are referencing in your course shell, link directly to that page rather than a site’s homepage or main host page.

Tutor can reference publisher content the same way it does any other linked page if that page is publicly viewable without a login or paywall. If the material sits behind student authentication (e.g., a textbook platform students have to log into), Tutor generally won’t be able to see it, since it can only read what’s on the open page itself. 

No. Tutor is programmed to act as a virtual study buddy, not to do the work for students. It’s explicitly instructed not to complete assignments. 

Yes, this is possible, but this points to an assignment design issue rather than a product limitation. Tutor is built to help students review and understand material, and that includes summarizing key concepts when asked. So, an assignment that asks students to essentially “summarize X” is something Tutor could complete for a student.  

The best mitigation is on the prompt side: assignments that ask students to apply, analyze, compare, or reflect on material (rather than just restate it) are much harder for Tutor to complete wholesale, since they require original synthesis rather than summarization.

No, Tutor will only use material that is set to be visible to students. It will not pull hidden materials or materials outside of a set availability (visibility) timeframe. For example, it won’t include material that you have set to appear to students at a future date. Do note that faculty can interact with Tutor in their course shell just like students do, but keep in mind it will include hidden materials when interacting with you because all content is visible to you.

When multiple sections are combined into a single merged course shell, the AI assistant indexes and scopes its knowledge base uniformly to the unified container. It treats the combined materials, modules, and schedule within that single merged shell as one unified source of truth for all enrolled students. If sections 2 and 3 are combined under section 1 and all learning material is set up in section 1 as normal, students in sections 2 and 3 will be able to use Lumi Tutor the same as students in section 1.

Lumi Feedback is designed to generate feedback from the assignment, rubric, submitted work and instructor settings, with the rubric as the primary determinant of a grade. Lumi Feedback addresses bias by keeping instructors in control and grounding AI feedback in the rubric, assignment requirements, and course context. AI-generated feedback is editable before it is shared with students, and the workflow supports consistent, fair communication across submissions rather than replacing instructor judgment.

As a further means of protecting students against grading bias, TCU’s Feedback tool has been given explicit custom instructions about evaluating the student only on the submitted work, disregarding student disclosures about their identity within the submission, and disregarding any instructor instruction that asks it to adjust feedback or scoring based on student identity.