IT Services Knowledge - Pangram-Canvas Integration

What Is Pangram?

Pangram is an AI detection and plagiarism detection tool that identifies potential AI-generated content and plagiarism in text. It analyzes text patterns to distinguish between human writing, fully AI-generated content, and human writing that has been edited or enhanced using AI tools. Pangram is not intended to be used on non-text documents like problem sets including math, chemical formulas, or on image, diagram, or graph-heavy work. For more about Pangram, see How does Pangram work?

The University has an enterprise license for Pangram and has secured the University’s data privacy and security. The data submitted to Pangram is not used to train their model, the data can be deleted on demand, and any student data submitted is FERPA-protected. For more details, review the Frequently Asked Questions.

Get Started with Pangram in Canvas

There are two ways of using Pangram within Canvas. In both cases, SpeedGrader will display a report of the likelihood that the student's submission is AI-generated. However, the range of tools available to you in SpeedGrader will vary depending on the route you choose.

If you create a Canvas Assignment of type Online Submission and add Pangram as a Document Processing App, you will have full capabilities to annotate student submissions in SpeedGrader.
If you create an Assignment of type External Tool and choose Pangram, you will not be able to annotate student submissions in SpeedGrader.

For a step-by-step explanation of each method, please click on the appropriate link.

Create a Pangram Assignment with Annotation Capabilities (Document Processing App)
Create a Pangram Assignment without Annotation Capabilities (External Tool)

Create a Pangram Assignment with Annotation Capabilities (Document Processing App)

To create a Pangram Assignment with annotation capabilities in SpeedGrader:

Open your course in Canvas and go to Assignments.
Create a new assignment or edit an existing one.
Under Submission Type, select Online and ensure that File Uploads is checked. (You may also choose other options, such as Text Entry, if you wish.)
Scroll down until you see the button Add Document Processing App and click it.
In the dialog box that appears, click Pangram - Asset Processor.
In the following dialog box, you can choose whether you wish your students to be able to view their AI reports from Pangram. Once you have made this choice, click Add Pangram to Assignment.
Once the dialog box closes, you will be informed that Pangram is now installed. You can then finish setting up the assignment and publish it as normal.
When a student uploads a file submission and you view it in SpeedGrader, you will see the full range of annotation tools normally found there, in addition to a report on the likelihood of AI use. See below View Pangram Results in SpeedGrader and Interpret Pangram Results.

Create a Pangram Assignment without Annotation Capabilities (External Tool)

To create a Pangram Assignment without annotation capabilities in SpeedGrader:

Open your course in Canvas and go to Assignments.
Create a new assignment or edit an existing one.
Under Submission Type, select External Tool.
Click Find, then select Pangram from the list of available tools.
Configure your preferred options (text entry, file upload, or both) and save.
Once enabled, Pangram will automatically scan submissions for that assignment and display results on a separate tab in SpeedGrader.

View Pangram Results in SpeedGrader

Open a student's submission in SpeedGrader.
Look for the Pangram AI Analysis tab on the right side of the submission view.
The panel displays:
1. An overall AI likelihood score (0-100%)
2. Highlighted segments showing which portions of the text are flagged as likely AI-generated
3. A distinction between fully AI-generated text and AI-assisted text (i.e., human writing that was rewritten or paraphrased using AI)
Click any highlighted segment to see the confidence level for that specific passage.
You can also view the plagiarism result, if any, by clicking View Plagiarism at the bottom.
The resulting box will list the web source(s) Pangram believes the text was taken from.

Submit Files Directly to Pangram (Outside Canvas)

If you have a document not submitted through Canvas, you can upload it directly at pangram.com. Supported formats: PDF, DOCX, RTF. For more information on how to use the pangram.com website, see Check for AI Use or Plagiarism with the Pangram Web Tool.

Interpret Pangram Results

Score	Interpretation
0-20%	Likely human-written
21-59%	Uncertain — review highlighted segments before drawing conclusions
60-100%	High likelihood of AI-generated or AI-assisted content

Important: Pangram scores are one data point, not a verdict. Before taking any action, review the flagged segments, consider the context (e.g., a non-native English speaker's writing may read differently), and speak with the student directly. Do not use a Pangram result alone as the basis for an academic integrity referral.

The CCTL provides guidance on Navigating Academic Integrity in the AI Landscape on their website, addressing issues such as assessment design and communicating expectations, in addition to Pangram.

Supported Languages

Pangram supports detection in multiple languages. Please see their page What languages does Pangram work on? for details on which languages are currently supported. To learn about how they tested, review their language testing methodology on their blog.

Privacy and Data Handling

All submissions are processed in compliance with FERPA.
Student data is not used to train Pangram's or any other third-party AI models.
Submissions are encrypted in transit and at rest.

Support and Resources

Resource	Link
UChicago IT Help Desk	uchicago.edu/ithelp
Canvas at UChicago	canvas.uchicago.edu
Academic Technology Solutions	academictech.uchicago.edu
Academic Integrity Office	college.uchicago.edu/student-services/academic-integrity-policies

Pangram-Canvas Integration