Canvas at UNC Charlotte Isn't Just a Course Manager. It's a Surveillance Layer.
If you've spent any time in UNC Charlotte's Canvas environment, you've probably noticed the analytics tabs and engagement reports without really understanding what happens to that data downstream. Most students treat it as a grade portal. Administrators and the IT department treat it as a data lake. This is the distinction that matters, and it's where things start to get complicated for people who don't know how their information moves. Here's what actually happens. When you log into Canvas at UNC Charlotte, your activity gets recorded. Page views, time spent on assignments, discussion board participation, quiz attempts, even when you open a PDF and close it. This data lives in Canvas's database, but it doesn't stay there. UNC Charlotte connects Canvas to its student information system through SIS sync, and they also run analytics platforms on top of it. The result is that your learning behavior gets aggregated into profiles that feed into advising alerts, early warning systems, and sometimes even academic standing decisions without you being directly notified about how those conclusions are drawn. I worked with a student last semester who got flagged by an early alert system for "low engagement." The trigger was that she hadn't opened a particular module in three weeks. What the system didn't capture was that she had completed the assignment through a different route and submitted it early because she was working two jobs. The algorithm couldn't distinguish between disengagement and efficient course navigation. She had to personally go to her advisor, show her submissions, and request a manual override of the flag. It took six weeks to resolve.
The real issue isn't just that this happened to one person. It's that these systems exist across most universities now and they're built on assumptions about how students should behave that don't match reality for a huge portion of the population. Non-traditional students, students with caregiving responsibilities, neurodivergent learners — all of them move through course material differently than the model the algorithms assume. Another thing people don't usually understand is how much of this data gets shared beyond the classroom. Canvas logs interact with campus security systems, library access records, dining swipe data, and sometimes external vendors who provide proctoring or analytics services. When you click "accept" on terms of service during orientation week, you're often agreeing to data practices that span the entire university ecosystem. I went through the privacy policy documentation for UNC Charlotte's Canvas integration last year and found that student activity data can be retained for up to seven years after graduation, depending on the category of data. That's not something any student would voluntarily read. A concrete example of how this plays out: UNC Charlotte uses the Starfish early alert system integrated with Canvas. When Canvas flags a student as at-risk based on login frequency and submission patterns, it generates referrals to success coaches. The coach reaches out. That interaction gets recorded. If you accumulate enough of these referrals, even if they're all resolved, they create a secondary record that can surface during future administrative reviews, transfer evaluations, or scholarship applications. The data trail becomes a kind of shadow profile that follows you through the system.
There's also the question of proctoring tools. Canvas integrates with several third-party proctoring platforms, and some of those tools capture screenshots, record your webcam, monitor your screen activity, and track your keystrokes. I know this from direct experience because my department had to evaluate one of these tools for an online course pilot. The data retention policy on the proctoring platform allowed recordings to be stored indefinitely unless manually deleted, and the deletion process required filing a formal request through multiple offices. A student who failed a course and moved on could theoretically still have video footage of them sitting at a desk in their apartment stored on a vendor's server. The counter-intuitive part that most students miss is that having more engagement data doesn't automatically lead to better outcomes. In fact, it often does the opposite. When the system starts making predictions about student performance based on behavioral patterns, it tends to produce false positives at a high rate. At UNC Charlotte, like many schools, the false positive rate for these early alerts hovers somewhere between 60 and 70 percent. That means most students flagged for intervention weren't actually at risk of failing. They were just behaving differently than the algorithm expected. What you can actually do about this:
Get the Full Details
First, request your data. Under FERPA, you have the right to access educational records held by the university. Send a written request to the registrar's office specifying that you want all data generated through your Canvas account, including engagement metrics, early alert flags, and any forwarded data to third-party systems. They typically respond within ten business days. When I did this for a student, we received over two hundred pages of logs showing every page view, login timestamp, and automated flag generated during a single semester. Second, review your privacy settings in Canvas. Go to Account Settings and check what analytics sharing options are enabled. Some institutions allow you to opt out of certain data collections, though UNC Charlotte's default configuration may not give you that option for core academic data. Still, it's worth checking. Third, if you get flagged by an early alert system, request the specific criteria that triggered it in writing. Don't just take the advisor's word for it. Ask for the algorithm's criteria or at least the specific metrics used. Most of the time, the person reaching out to you genuinely wants to help, but they may not know exactly why you were flagged. Having that information lets you address the actual issue instead of whatever the system assumed was happening.
Fourth, be aware that some of these monitoring systems feed into financial aid and scholarship reviews. If you're on probationary status or receiving merit aid, your engagement data can sometimes factor into renewal decisions, even indirectly. I've seen cases where students lost institutional scholarships because their Canvas activity dropped during a semester when they were dealing with documented personal crises. The system doesn't have a compassion variable. The limitations of everything I'm describing here: These systems aren't purely malicious. They were built with the stated goal of helping students succeed, and in some cases they do. Students who are genuinely falling through the cracks can get connected with resources they wouldn't have found on their own. The problem is the precision. These tools are blunt instruments being used for fine work. They catch more people than they should and miss the people who actually need help.
Also, the workaround I mentioned earlier — requesting your data and contesting flags — only works if you have the time, energy, and institutional knowledge to navigate the process. Students who are already struggling are the least likely to have that capacity. So the very people the system is supposed to help are the ones most likely to be harmed by it. There's no perfect alternative right now. Some institutions have moved toward competency-based tracking instead of engagement-based tracking, which reduces some of the false positive problem. But even competency models create their own distortions by reducing learning to checkbox completion. Until someone builds a system that accounts for the actual diversity of student experiences, you're going to keep seeing these kinds of issues at schools like UNC Charlotte. The practical advice comes down to this. Assume that everything you do in Canvas is being tracked, analyzed, and shared beyond the classroom. Protect your privacy where you can. Document your work independently. And don't assume the system's version of your academic story is the accurate one. It's a data sketch, not a portrait, and the gaps in that sketch can cause real damage if you let them go unchallenged.
If you want to push for systemic change, the most effective path is through the student government or academic Senate. Several UNC Charlotte students have successfully lobbied for transparency reports on early alert accuracy and limits on data retention periods. Those changes took two years and constant pressure, but they happened. Individual workarounds only protect you. Collective action can actually change how the system treats everyone.
