No test results can be considered legitimate unless the testing environment creates an even playing field that favors no students over others. Unfortunately, the algorithmic bias of off-the-shelf AI-based remote proctoring solutions puts some students at a disadvantage before the test has even begun. This is particularly true when these proctoring systems rely on general-purpose datasets that do not reflect the student population or the testing environment, raising the risk of false positives.
Eliminating built-in bias requires a remote proctoring approach that recognizes and respects the diversity of the student population. Equity and inclusion lie at the heart of Rosalyn’s human-in-the-loop (HITL) proctoring system, which is based on purpose-built, constantly improving AI. Rosalyn’s machine learning algorithms rely on an ever-expanding dataset that represents a full spectrum of genders, ages, skin tones, ethnicities, and abilities, continually sharpening the system’s ability to treat all students fairly. There is also a second layer of protection, as all potential violations must be accepted or rejected by a human proctor.
Students at universities around the world are petitioning their schools to stop proctoring services from violating privacy—one of the most challenging online proctoring disadvantages to overcome. Their concerns relate to everything from biometric data collection to overbroad access to students’ computers to serious cybersecurity vulnerabilities. Most proctoring systems also leave students in the dark about which data is collected during the test-taking process, leaving no opportunity for informed consent.
Protecting the privacy and dignity of students must be a priority for online proctoring companies. Rosalyn collects only the data required to verify the student’s identity and ensure all test-takers are treated equitably. Students can clearly understand which data is being collected, how long it will be retained, and how it will be used. Data is never sold or otherwise shared with third parties.
In the past, whether students took a test remotely was usually the test-taker’s decision. That changed quickly in 2020 when many schools began relying almost exclusively on online instruction. One result of mandatory online exams was a sharp uptick in students expressing dissatisfaction with remote proctoring solutions. This negative perception can seriously damage trust in educational institutions and deeply affect the student experience.
Changing perceptions of remote proctoring requires being transparent about how remote proctoring works and how the right technology can benefit students. When students feel comfortable with online proctoring systems, schools can rebuild trust and students can perform at their best.
Students, educators, and advocacy groups have all raised the alarm about the discrimination some remote proctoring systems perpetuate against disabled students. Not only do students with disabilities face more challenges accessing the technology required for online test-taking, the systems used to monitor tests often incorrectly flag potential violations, as core features like keystroke logging, facial recognition, and movement tracking have not been designed with disabilities in mind.
The solution is to use a remote proctoring solution that is trained using datasets that include students with a range of disabilities, minimizing the risk of false flagging. AI should work in tandem with human proctors to ensure that disabled test-takers enjoy the same unbiased testing environment as other students.
Some remote proctoring systems collect significant personal information about students, capturing images, sounds, and movements throughout the testing process. Not only can this raise serious concerns about privacy, it can also leave students keenly aware that their every move is being judged, whether by a machine or by a human watching and listening from a remote location.
Many students are also worried about having no recourse when proctoring is performed by automated systems. The potential for systems to unjustly flag events when no violations have actually occurred is a serious concern that can have a profound impact on students’ academic careers. Relying solely on technology without human oversight can greatly increase stress and sow distrust in exam results.
Considering these and other online proctoring disadvantages, it’s no wonder so many students feel more anxious about taking a remote exam than they do about tests administered in a classroom setting. Putting students at ease with the remote proctoring process by limiting intrusiveness, protecting privacy, and putting final decision-making in the hands of humans is the focus of Rosalyn’s HITL proctoring approach.
The technology required for online proctoring can put students at significant disadvantage. A recent study by researchers at the Amsterdam Center for Learning Analytics found that students taking an online exam that was proctored remotely were six times more likely to experience problems with test-taking than students taking the same test in a classroom. These difficulties can arise from:
Reducing the technology demands of remote proctoring is essential to making systems more accessible to all students. Rosalyn extends its commitment to student fairness by anticipating and mitigating any potential technical obstacles students may encounter while taking a remote exam. This includes featuring a simple check-in process, compatibility with multiple platforms, and low bandwidth and connectivity requirements.
Some systems charge students per-test fees, creating significant barriers to education for some students and perpetuating economic inequality. However, there can be even greater costs associated with online test administration.
Fairness to students includes an equitable exchange when taking an online exam. Some proctoring solutions demand more from students than they provide them in return, such as access to significant personal information. This is one of the most important online proctoring disadvantages proctoring vendors often fail to appreciate; for many students, surrendering their privacy and dignity simply isn’t worth it.
Rosalyn is committed to respecting students and ensuring that they receive fair value for all the costs they incur as a result of remote proctoring. This includes minimizing the technological barrier to entry, remaining cost-effective for educational institutions, and not demanding access to sensitive information.
Many instructors are reluctant to rely on remote testing until they can be convinced that the process is as equitable and effective as in-classroom testing is. They remain unconvinced of the ability of automated proctoring systems to prevent test violations without penalizing students or affecting the accuracy of the test results. More importantly, they are often deeply concerned about the potentially destructive impact on student dignity and trust in the exam process.
As educators become more familiar with cutting-edge solutions to online proctoring disadvantages, they are better able to adapt their tests and the test-taking process to overcome the shortcomings of remote proctoring. This includes selecting remote invigilation systems that prioritize fairness and transparency to students, are designed to learn and improve over time, and marry automation with human wisdom. With these tools, instructors and institutions can lay the foundation for student success.
This comprehensive guide explores the evolution of proctoring services, delving into the intricacies and comparisons of different AI proctoring models.
As artificial intelligence (AI) takes center stage across industries, it brings misconceptions, especially in remote exam proctoring—a field accelerated by COVID-19.
Welcome to the future of cheating, where AI isn't just an ally but an accomplice. Discover how Rosalyn.ai is combating this threat.