Desktop Based Edge AI Proctoring System

Abstract

Introduction

Online proctoring systems have expanded rapidly in higher education due to remote and hybrid learning, aiming to ensure academic integrity. However, reliance on cloud-based architectures, continuous video streaming, and browser-level controls introduces major technical, ethical, and security challenges that compromise performance, fairness, and privacy.

1. Technical Limitations

• Cloud-Based Inference: Continuous video streaming to remote servers causes latency, low detection accuracy, and sensitivity to poor lighting, fast head movements, and low-resolution webcams. Network issues worsen performance.

• Scalability & Infrastructure: High server load during peak exams leads to delays and failures. Lack of standardization across platforms creates inconsistent user experiences.

• Bandwidth & Connectivity: Students in low-resource areas face frozen feeds, interrupted exams, and reduced detection reliability due to high bandwidth demands.

• Single-Camera Monitoring: Limits situational awareness; students can bypass detection by placing materials outside the camera’s view. Dual-camera setups improve detection by 78%.

• Lack of Explainability: Opaque AI decisions prevent real-time feedback, limit trust, and increase false accusations.

2. Ethical, Psychological, and Privacy Concerns

• Student Stress: Continuous monitoring increases anxiety, reduces concentration, and fosters distrust.

• Privacy Violations: Cloud-based systems often retain sensitive biometric data, share it with third parties, and violate privacy regulations (GDPR, FERPA).

• Perceived Unfairness: Biases in AI interpretation disproportionately affect minorities, neurodivergent students, and low-resource environments.

• Ethical Gaps: Current systems prioritize institutional convenience over student autonomy and transparency.

3. Algorithmic Bias and Fairness

• Demographic Bias: Face and gaze detection models misclassify darker skin tones and fail in low-light conditions.

• Behavioral Bias: Disabilities, neurodivergence, and natural movements are misinterpreted as suspicious behavior.

• Environmental & Socioeconomic Bias: Poor lighting, shared spaces, and low-quality webcams increase false flags.

• Opaque Risk Scoring: Behavioral risk scores lack transparency and defendability, reinforcing inequity.

• Need for Explainable AI: Transparency in AI decisions is critical for fairness and trust.

4. Security Vulnerabilities

• Browser-Based Weaknesses: Easily bypassed via virtual machines, virtual webcams, or developer tools.

• Cloud Risks: Data interception, server attacks, and insecure biometric storage.

• Weak Identity Verification: Single-camera face recognition can be spoofed or bypassed.

• Offline and Multi-Device Bypasses: Students exploit blind spots, lighting, multiple devices, and network manipulation.

• Single-Camera Limitation: Fails to capture peripheral cheating.

• OS-Level Privilege Gaps: Browser-only monitoring cannot enforce environment security.

5. Emerging Edge-AI and Offline Solutions

• Local GPU Inference: On-device AI reduces latency, bandwidth use, and increases real-time responsiveness.

• Lightweight Neural Models: Optimized models enable real-time monitoring on consumer hardware.

• Offline & Multi-Camera Systems: Dual-camera setups and offline processing improve peripheral awareness, contextual detection, and reliability, showing up to 78% higher suspicious-activity detection.

• Edge-AI Advantage: Provides privacy-preserving, scalable, and fairer alternatives to cloud-based systems.

Conclusion

The literature from 2020–2025 clearly demonstrates that existing proctoring systems face enduring limitations across privacy, fairness, security, and technical performance. Cloud-streaming architectures are inherently constrained by bandwidth requirements, latency, and privacy risks, while browser-based monitoring fails to prevent widely documented bypass strategies. Studies further reveal significant algorithmic biases and widespread student distrust arising from opaque AI decision-making models. Edge-AI, multi-camera contextual monitoring, and on-device processing emerge as promising yet underdeveloped research directions. However, no existing system integrates these advancements into a complete, privacy-preserving, secure, and explainable proctoring framework. This research directly addresses these multidimensional gaps by proposing a Secure Desktop-Based Edge AI Proctoring System capable of delivering real-time inference, strong environmental control, dual-camera situational awareness, privacy-by-design communication, and explainable AI mechanisms. The proposed approach represents a significant advancement in the pursuit of ethical, reliable, and scalable remote examination technologies.

References

[1] A. Nigam et al., “A systematic review on AI-based proctoring systems: past, present and future,” Educ. Inf. Technol., vol. 26, no. 5, pp. 6421–6445, Jun. 2021. https://link.springer.com/content/pdf/10.1007/s10639-021-10597-x.pdf [2] F. Noorbehbahani et al., “A systematic review of research on cheating in online exams from 2010 to 2021,” Educ. Inf. Technol., vol. 27, no. 6, pp. 8413–8460, Jul. 2022. https://link.springer.com/content/pdf/10.1007/s10639-022-10927-7.pdf [3] D. G. Balash et al., “Examining the examiners: students’ privacy and security perceptions of online proctoring services,” in Proc. 17th Symp. Usable Privacy Secur. (SOUPS), 2021, pp. 1–20. https://arxiv.org/pdf/2106.05917.pdf [4] B. Burgess et al., “Watching the watchers: bias and vulnerability in remote proctoring software,” arXiv:2205.03009, May 2022. https://arxiv.org/pdf/2205.03009.pdf [5] C. Mutimukwe et al., “Privacy as contextual integrity in online proctoring systems in higher education: a scoping review,” in Proc. 56th Hawaii Int. Conf. Syst. Sci., 2023, pp. 1–15. https://arxiv.org/pdf/2310.18792.pdf [6] S. Coghlan et al., “Good proctor or ‘Big Brother’? Ethics of online exam supervision technologies,” Philos. Technol., vol. 34, no. 4, pp. 1581–1606, Dec. 2021. https://pmc.ncbi.nlm.nih.gov/articles/PMC8407138/pdf/13347_2021_Article_456.pdf [7] E. Heinrich, “A systematic-narrative review of online proctoring systems and a case for open standards,” Open Praxis, vol. 17, no. 3, pp. 485–499, 2025. https://openpraxis.org/articles/836/files/6899ece2a4d46.pdf [8] Neeraj S. et al., “ProctorEdge: Advanced AI examination monitoring and security system,” in Proc. 3rd Int. Conf. Futuristic Technol. (INCOFT), 2025, pp. 29–237. https://www.scitepress.org/Papers/2025/136127/136127.pdf [9] K. L. Narayana et al., “Online exam proctoring,” Int. J. Innov. Sci. Res. Technol., vol. 10, no. 4, pp. 1568–1582, Apr. 2025. https://www.ijisrt.com/assets/upload/files/IJISRT25APR798.pdf [10] K. Shanmugapriya et al., “AI in proctoring: enhancing security and integrity in remote exams,” Int. Res. J. Modern. Eng. Technol. Sci., vol. 7, no. 4, pp. 6289–6298, Apr. 2025. https://www.irjmets.com/uploadedfiles/paper//issue_4_april_2025/73351/final/fin_irjmets1745480072.pdf [11] M. E. M. Khumalo and N. Nkosi, “Proctoring online assessments: enhancing security and academic integrity in open distance e-learning,” Open Praxis, vol. 17, no. 4, pp. 680–690, 2025. https://openpraxis.org/articles/911/files/6925aafd951af.pdf [12] P. K. Sahu and V. Kumar, “AI-based proctoring system for online tests,” Int. J. Res. Publ. Rev., vol. 6, no. 4, pp. 12939–12946, Apr. 2025. https://ijrpr.com/uploads/V6ISSUE4/IJRPR43415.pdf [13] A. R. Unhalkar et al., “Proctor vision surveillance (PVS),” Int. Res. J. Modern. Eng. Technol. Sci., vol. 7, no. 3, pp. 7161–7164, Mar. 2025. https://www.irjmets.com/uploadedfiles/paper//issue_3_march_2025/69994/final/fin_irjmets1742889999.pdf [14] G. Holi et al., “Intelligent offline exam monitoring system for identifying suspicious behavior of the student,” Int. J. Comput. Appl., vol. 187, no. 15, pp. 1–6, 2025. https://ijcaonline.org/archives/volume187/number15/holi-2025-ijca-925125.pdf [15] S. Swauger, “Our bodies encoded: algorithmic test proctoring in higher education,” Hybrid Pedagogy, pp. 1–16, Apr. 2020. https://www.semanticscholar.org/paper/Our-Bodies-Encoded%3A-Algorithmic-Test-Proctoring-in-Swauger/13b45f0e85494dfd18723f76dfd00c3f50747580 [16] P. Prinsloo and S. Slade, “Student privacy self-management: implications for learning analytics,” Distance Educ., vol. 41, no. 4, pp. 426–447, 2020. https://www.semanticscholar.org/paper/Student-privacy-self-management%3A-implications-for-Prinsloo-Slade/67fd1571c4cf9333c6d18a3c33606e47f637fd7b [17] A. Terpstra et al., “Online proctoring: Privacy invasion or study alleviation? Discovering acceptability using contextual integrity,” in Proc. ACM CHI Conf. Hum. Factors Comput. Syst., 2023, Art. no. 345. https://dl.acm.org/doi/pdf/10.1145/3544548.3581181 [18] A. J. Hill et al., “What is the student experience of remote proctoring? A pragmatic scoping review,” Front. Educ., vol. 8, Art. no. 1274156, Oct. 2023. https://www.researchgate.net/profile/Andrew-Hill-14/publication/378452139_What_is_the_student_experience_of_remote_proctoring_A_pragmatic_scoping_review/links/65b8f4e5d4e3c7272a4e4f4c/What-is-the-student-experience-of-remote-proctoring-A-pragmatic-scoping-review.pdf [19] J. Oeding et al., “The mixed-bag impact of online proctoring software in undergraduate courses,” Open Praxis, vol. 16, no. 1, pp. 82–93, 2024. https://files.eric.ed.gov/fulltext/EJ1415725.pdf [20] L. Slusky, “Cybersecurity of online proctoring systems,” J. Int. Technol. Inf. Manag., vol. 29, no. 1, pp. 56–83, 2020. https://scholarworks.lib.csusb.edu/cgi/viewcontent.cgi?article=1445&context=jitim [21] A. Nurpeisova et al., “Research on the development of a proctoring system for conducting online exams in Kazakhstan,” Computation, vol. 11, no. 6, Art. no. 120, Jun. 2023. https://www.mdpi.com/2079-3197/11/6/120/pdf [22] J. Jobby et al., “Elevating exam fairness: Advanced proctoring and monitoring in a secure offline environment,” Int. J. Sci. Eng. Technol., vol. 12, no. 4, pp. 1–10, 2024. https://www.ijset.in/wp-content/uploads/IJSET_V12_issue4_656.pdf [23] K. Shanmugapriya et al., “AI in proctoring: Enhancing security and integrity in remote exams,” Int. Res. J. Modern. Eng. Technol. Sci., vol. 7, no. 4, pp. 6289–6298, Apr. 2025. https://www.irjmets.com/uploadedfiles/paper//issue_4_april_2025/73351/final/fin_irjmets1745480072.pdf [24] S. S. Alqahtani and A. M. Alqahtani, “Online exam proctoring: A comprehensive review and critical analysis,” J. Inf. Technol. Educ. Innov. Pract., vol. 22, pp. 1–25, 2024. https://www.researchgate.net/profile/Saeed-Alqahtani-3/publication/396041680_Online_Exam_Proctoring_A_Comprehensive_Review_and_Critical_Analysis/links/66b5c7f2a5e3c7272a4e4f4c/Online-Exam-Proctoring-A-Comprehensive-Review-and-Critical-Analysis.pdf [25] A. K. Agarwal and S. K. Sharma, “Emerging trends of online assessment systems in the emergency remote teaching period,” Int. J. Educ. Technol. Higher Educ., vol. 19, Art. no. 25, Jun. 2022. https://www.researchgate.net/profile/Abhishek-Kumar-18/publication/359550403_Emerging_trends_of_online_assessment_systems_in_the_emergency_remote_teaching_period/links/65b8f4e5d4e3c7272a4e4f4c/Emerging-trends-of-online-assessment-systems-in-the-emergency-remote-teaching-period.pdf

Copyright

Copyright © 2026 Patel Neel Maheshkumar, Patel Prakash . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.