Topics Map > Ed Tech Tools > CANVAS > Discussions
Canvas Discussions - Discussion Summarization (Feature Preview) Best Practices, Known Issues
About the Discussion Summarization Feature Preview
The Discussion Summarization feature in Canvas LMS uses generative AI to provide instructors with summaries of key points, questions, and ideas from discussion threads. This feature is intended to save educators time and make discussions more manageable in large-format classes.
- Discussion Summaries FAQ (Instructure Community)
- Discussion Summaries Change Log (Instructure Community)
Ethical Considerations
Since the Discussion Summarization feature utilizes AI, it's crucial to address the ethical implications associated with its use.
- Privacy: It's essential to ensure that the AI tool adheres to privacy regulations and safeguards student data.
- Instructure has published a "Nutrition Facts" article that describes the underlying technology and the way it handles data. All data submitted to Canvas LMS is subject to the policies which govern handling of student and staff data by the Universities of Wisconsin.
-
- Transparency: Instructors should inform students that an AI tool is being used to summarize discussions and explain how it is being used.
- Over-reliance: Instructors should be mindful of over-reliance on the AI tool and maintain human oversight in interpreting and utilizing the summaries.
Best Practices for Using Discussion Summarization in Canvas LMS
To effectively use the Discussion Summarization feature in Canvas LMS, instructors can consider the following best practices:
- Access the summarize button: Once enabled, instructors will find a "Summarize" button at the top of the Discussions page.
- Use prompts: Utilize the prompt feature to guide the AI in summarizing specific aspects of the discussion. Canvas provides examples of effective prompts to help instructors get started
- Refine and regenerate summaries: If the initial summary is not satisfactory, instructors can refine it by providing more specific prompts or regenerate it for a different perspective.
- Provide feedback: Use the thumbs up/down feature to provide feedback on the quality of the summary, which can help improve the AI's performance over time.
- Integrate summaries into instruction: Share the summary with students by copying and pasting it into a discussion reply or an announcement. Edit the summary as needed to provide context or further explanation.
- It's important to remember that the Discussion Summary is only visible to the instructor, allowing for personalized feedback and targeted discussions without publicly revealing individual student struggles
- It's important to remember that the Discussion Summary is only visible to the instructor, allowing for personalized feedback and targeted discussions without publicly revealing individual student struggles
- Set clear expectations: Provide clear guidelines for discussion participation and grading, ensuring students understand the purpose and value of the discussions.
- Facilitate ongoing discussion: Use the summary to identify key themes and questions that require further discussion or clarification. Encourage students to engage with the summary and each other.
Alternative Summarization Methods
While the AI-powered Discussion Summarization feature offers a convenient way to analyze discussions, instructors can also consider alternative methods, especially if the AI tool is unavailable or not producing satisfactory results. One such method is using word clouds, which visually represent the most frequent words in a text. Instructors can copy and paste student posts into a word cloud generator to create an image that highlights the key themes and topics discussed. This can be a useful way to engage students and provide a quick overview of the discussion.
Known Issues and Problems
Instructors should be mindful of these potential issues and consider them when interpreting and utilizing the generated summaries. Providing feedback on inaccurate or misleading summaries can help improve the AI's performance over time.
| Issue | Description |
|---|---|
| Generation limit | There is a limit of 25 summary generations per user per day. Once this limit is reached, users must wait until the next day to generate more summaries |
| Loss of user-refined summaries | If a user customizes the summary by inputting their own text, the refined summary might be lost when the page is refreshed |
| Disabling summaries | When the "Disable Summary" button is selected, it currently disables the summary for all instructors in the course. However, it might be more beneficial if it only disabled the summary for the specific user who clicked the button |
| Conflating information | The AI might sometimes conflate information from different sources, including other Canvas institutions or external websites. This can lead to inaccurate or misleading summaries. Instructors should carefully review and verify the summaries, especially when dealing with specific details or potentially sensitive information. |
| Odd responses | The AI might occasionally generate odd or unexpected responses, particularly when asked to identify specific information or analyze images within the discussion |
| Repetition | If similar points are made multiple times in the discussion, the extractive summary might include redundant information |
More Information on Discussion Summarization
- Discussion Summaries (Feature Preview) - Instructure Community
- Canvas LMS: Discussion Summaries
- Discussions Redesign & AI Disscussions | Instructure
For Further Assistance
- Visit the Instructure Community for the Discussion Summaries Feature Preview to share your feedback and engage with other instructors using the feature.
- Contact UWM Canvas Support with your questions
Note: This article was drafted using Google Gemini Advanced using the "1.5 Pro with Deep Reasoning" model, and Microsoft Copilot Chat; then revised for brevity and audience. The content was reviewed, verified, and edited by a Canvas Administrator.