Over the past few months, students from all over the country have reached out to EFF and other advocacy organizations because their schools—including teachers and administrators—have made flimsy claims about cheating based on digital logs from online learning platforms that don’t hold up to scrutiny. Such claims were made against over a dozen students at the Dartmouth Geisel School of Medicine, which EFF and the Foundation for Individual Rights in Education (FIRE) criticized for being a misuse, and misunderstanding, of the online learning platform technology. Dartmouth ended that investigation and dismissed all allegations after a media firestorm. If your school is making similar accusations against students, here’s what we recommend.
Students Deserve the Evidence Against Them
Online learning platforms provide a variety of digital logs to teachers and administrators, but those same logs are not always made available to the accused students. This is unfair. True due process for cheating allegations requires that students see the evidence against them, whether that’s videos from proctoring tools, or logs from test-taking or learning management platforms like Canvas or Blackboard.
Schools should use technology to serve students, rather than using it as a tool to discipline them.
It can be difficult to know what logs to ask for, because different online learning platforms call this data by different names. In the case of Canvas, there may be multiple types of logs, depending on whether a student used the platform to take a test or access course materials while studying for it.
Bottom line: students should be given copies of any logs that are being cited as evidence of cheating, and any logs that may be exculpatory. It’s all too easy for schools to cherry-pick logs that only indicate possible misconduct. With course material access logs, for example, schools often only share (if they share at all) logs that indicate a student’s device accessed material that is relevant to the subject of the test, while dismissing logs that show access of materials that are less relevant, thus hiding evidence that the access was the result of an automated link between the device and platform. Any allegation should start with the student being shown everything that the administration has access to—and we’re calling on learning platforms like Canvas and Blackboard to give students direct access, too.
A sample log from Blackboard
Digital Logs Are Unreliable Evidence of Cheating
It’s important for both students and school officials to understand why digital logs are unreliable evidence of cheating. Course material access logs, for example, can only show that a page, document, or file was accessed by a device—not necessarily why or by whom (if anyone). Much like a cell phone pinging a tower, logs may show files being pinged by a device in short time periods, suggesting a non-deliberate process, as was the case with the access logs we saw from Dartmouth medical students. It can be impossible to know for sure from the logs alone if a student intentionally accessed any of the files, or if the pings happened due to delayed loading, or automatic refresh processes that are commonplace in most websites and online services.
Canvas, for its part, has stated multiple times that both test-taking logs and course material access logs are not reliable. According to the company, test-taking logs, which purport to show student activity during a Canvas-administered test, “are not intended to validate academic integrity or identify cheating for a quiz.” Similarly, logs that purport to show student access to class documents uploaded to Canvas, are also not accurate; as the company explains: “This data is meant to be used for rollups and analysis in the aggregate, not in isolation for auditing or other high-stakes analysis involving examining single users or small samples.”
Blackboard has so far not made any public statements on the accuracy of its logs, but when contacted, the company said they are working on a public disclaimer to avoid any misconceptions on the accuracy and use of this type of data. The company was clear that logs should not be used to allege cheating: “Blackboard does not recommend using this data alone to detect student misconduct, and further, when an inquiry is made by a client related to this type of investigation, Blackboard consistently advises on the possible inaccurate conclusions that can be drawn from the use of such data.” Both Canvas and Blackboard should be more transparent with their users about the accuracy of their logs. For now, it's imperative that educators and administrators understand the unreliability of these logs, which both companies have admitted, albeit not as openly as we would like.
Collaboration Between Students Can Be Key
If one student is being charged with cheating based on digital logs, it’s likely others are as well, so don’t be afraid of rallying fellow students. At Dartmouth medical school, collective activism helped individual students push back against the cheating allegations, ultimately forcing the administration to withdraw them. Dartmouth students accused of cheating worked together to uncover flaws in the investigation, then contacted advocacy organizations and the press, and held on-campus protests.
Sympathetic teachers and administrators may also be valuable resources when it comes to pointing out unreliable evidence and due process problems. It may also be helpful to reach out to a technologist where possible, given the technical expertise required to examine digital data. Even a school computer club may be able to offer assistance.
Surveillance Is Not the Solution
If a school is unable to use digital logs to prove cheating, the administration may consider adding even more invasive measures, like proctoring tools. But mandating more surveillance of students is not the answer. Schools should use technology to serve students, rather than using it as a tool to discipline them.
Disciplinary technologies that start by assuming guilt, rather than promoting trust, create a dangerous environment for students. Many schools now monitor online activity, like social media posts. They track what websites students visit. They require students to use technology on their laptops that collects and shares private data with third-party companies, while other schools have implemented flawed facial recognition technology. And many, many schools have on-campus cameras, more and more of which feed directly to police.
But these technologies are often dangerously biased, and profoundly ineffective. They rob students of the space to experiment and learn without being monitored at every turn. And they teach young people to expect and allow surveillance, particularly when a power imbalance makes it difficult to fight back, whether that monitoring is by a school, an employer, a romantic partner, or the government. This problem is not just a slippery slope—it’s a cliff, and we must not push an entire generation off of it. Privacy is a human right, and schools should be foundational in a young person’s understanding of what it means to live in a society that respects and protects human rights.
EFF’s Statement on the Use of E-Learning Platform Logs in Misconduct Allegations
If necessary, you may wish to forward to your teachers or administrators this blog post on the problems with using digital logs as evidence of academic misconduct. If course material access logs, specifically, are being cited against you, you may forward EFF’s statement below. While we cannot assist every student individually, we hope this will help guide schools away from improperly using digital logs as evidence of cheating:
As a nonprofit dedicated to defending digital privacy, free speech, and innovation—including in the classroom, our independent research and investigation has determined that there are several scenarios where course material access logs of e-learning platforms can be generated without any student interaction, for example, due to delayed loading on a device or due to automatic refreshing of webpages. Instructure, the company behind the e-learning platform Canvas, has publicly stated that their logs (both course material access logs and test-taking logs) are not accurate and should not be used for academic misconduct investigations. The New York Times, in their own investigation into Canvas access logs, found this to be true as well. Blackboard, as well, has stated that inaccurate conclusions can be drawn from the use of their logs. Any administrator or teacher who interprets digital logs as evidence that a student was cheating may very likely be turning false positives into accusations of academic misconduct.
Educators who seek out technical evidence of students cheating, whether those are through logs, proctoring apps, or other computer-generated techniques, must also seek out technical expertise, follow due process, and offer concrete routes of appeal to students. We urge universities to protect the due process rights of all students facing misconduct charges by ensuring basic procedural safeguards are in place to guarantee fairness. These include, among other things, access to the full suite of evidence—including evidence that might tend to exculpate the student—and sufficient technical guidance for factfinders to interpret the evidence marshaled against the student. Students should also have time to meaningfully prepare for any hearing. These safeguards are necessary to ensure a just and trustworthy outcome is reached.
