Purpose and Disclaimer
Student potential misuse of Artificial Intelligence (AI) is a prevalent concern for many instructors. This resource is intended to direct instructors to available data in Canvas and frame insights in terms of patterns and anomalies that may warrant additional attention should concerns about student AI misuse arise. Please note: Identified data cannot be used alone to determine student use of AI or allege academic misconduct. This resource serves merely as a starting point on what data is available and where in Canvas related to student course activity.
Data Sources in Canvas
Individual Student Quiz Activity
How To Find It
- How do I view a quiz log for a student? (Classic Quizzes)
- How do I use the Moderate page in New Quizzes? (New Quizzes)
What To Look For
Examine time spent per question and overall quiz attempt duration. Intriguing data may include:
- Unusually rapid quiz attempt completion
- Remarkable changes to answer submission timing patterns from previous quizzes e.g. previous quiz answer submission times varied in length while current answer times were uniformly quick despite variations in question type and complexity
- Periods of inactivity coinciding with simultaneous submission of multiple answers
Look at the quiz log for breaks in quiz attempts. The log will indicate when a student stops viewing the quiz-taking page and when they resume viewing the quiz-taking page.
Aggregate Quiz Data
How To Find It
- Once I publish a quiz, what kinds of quiz statistics are available? (Classic Quizzes)
- How do I view reports for a quiz in New Quizzes? (New Quizzes)
What To Look For
Use Average Time and Average Score data to evaluate the student’s performance in comparison with peers. Intriguing data may include:
- Correctly answering questions when all or most other students answered incorrectly
- Answering complex or difficult questions much quicker than all or most other students
Student Course Access Report
How To Find It
How do I view the course access report for an individual user?
What To Look For
Use the Course Access Report to gain insight related to student interaction with course content in relation to their performance in the course. Intriguing data may include:
- Low or no engagement with instructional materials when paired with high achieving assessment and/or assignment submissions
Course Analytics
How To Find It
How do I view Course Analytics in a course as an instructor?
What To Look For
Use the Course Activity Report to identify patterns or shifts in the frequency and duration of student interaction with content. Intriguing data may include:
- Low or no engagement with instructional materials when paired with high achieving assessment and/or assignment submissions
- A change in typical levels of course activity coinciding with observed changes in assignment quality or style
- Brief interaction time with assignments and assessments coinciding with a lack of interaction with supporting instructional material
- Lower than average course interaction coinciding with higher-than-average course performance
Get All the Information
The available data in Canvas provides a portion of a larger story. While information may be suggestive of certain student behaviors, there can be other explanations for the activity seen. It is important to refrain from making judgments without additional information or consideration.
- Growth is expected, but a sudden change may be suspicious. Look for anomalies versus shifts over time. Compare the student’s writing style, answering patterns, and their academic performance across multiple assignments and assessments.
- Does the effort match the outcome? Student participation data in Canvas is an important metric, but it is not fully inclusive of student effort. Canvas cannot capture student work done outside of the system. In-person participation, tutoring, independent reading, and studying are important factors. Further, working on a response in a draft document outside of Canvas that is then copied in would not be captured. AI usage is not the only possible explanation for low interaction data in Canvas coinciding with high student performance.
- Is there a technical explanation for the data? Unusually quick answer submissions or submissions of multiple quiz answers at once may be indicative of AI, but it may also be the result of internet connectivity issues and Canvas safeguards against lost work.
- Interruptions happen. Periods of inactivity or multiple breaks in viewing the quiz-taking page may have a reasonable explanation.
When the data casts doubt on student behavior, begin with broad questions and explore other alternatives. Should you decide to move forward with a discussion with the student, ask questions that provide an opportunity for the student to share an unusual experience that may have happened during a quiz or provide insight on how they completed an assignment. They may offer the explanation you seek. Ask additional questions if needed and be sure to follow any Academic Misconduct Procedures.
Instructors are encouraged to maintain transparent expectations and dialogue regarding AI. For more information on setting expectations for generative AI use, assignment design considerations, academic misconduct, AI syllabus statements, and CASL workshops on AI, please refer to Artificial Intelligence (AI) and Teaching.