EUVIMEDEUVIMED entdecken
Crossref

Online Examination Security: A Multi-Layered Threat Framework and AI-Based Proctoring Architecture for Academic Integrity

Rushit Trivedi

Dandao Xuebao/Journal of Ballistics · 21.05.2026 · 38 (1) · 418-427

DOI: 10.52783/dxjb.v38.291 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Dandao Xuebao/Journal of Ballistics, online erschienen am 21.05.2026. Verfasst von Rushit Trivedi.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The rapid migration of academic assessment to online platforms, accelerated by the COVID-19 pandemic and sustained by the global expansion of distance education, has introduced a complex and evolving landscape of cybersecurity threats that fundamentally undermine academic integrity and the validity of educational credentials. Online examinations are now targeted by a diverse array of threat actors employing sophisticated techniques: session hijacking, man-in-the-middle (MITM) attacks, credential stuffing, AI-generated answer injection, deepfake identity substitution, screen-sharing collusion networks, and coordinated examination content harvesting operations. Existing e-proctoring solutions suffer from critical limitations: high false positive rates that disproportionately disadvantage marginalized students, inadequate network-layer security, privacy intrusions without commensurate security benefit, and a fundamental inability to detect AI-assisted cheating via Large Language Models (LLMs). This paper presents ExamShield-AI, a comprehensive multi-layered threat framework and AI-powered proctoring architecture designed to address the full spectrum of online examination threats while maintaining student privacy, examination fairness, and accessibility. ExamShield-AI integrates seven security layers: network traffic analysis, continuous multi-modal biometric verification, behavioral anomaly detection, AI-assisted answer analysis, secure browser enforcement, real-time incident response, and post-examination forensic audit. The system is evaluated across three real-world university deployment scenarios encompassing 47,280 examination sessions from 2023–2025, demonstrating 94.7% cheating detection accuracy with a false positive rate of only 1.84%, representing a 61.3% reduction in false positives compared to leading commercial proctoring platforms. The framework further proposes a comprehensive Online Examination Threat Taxonomy (OETT) classifying 28 distinct threat vectors across four attack surfaces, providing a foundational reference for academic institutions, examination bodies, and cybersecurity researchers.

Zitieren

APA

Trivedi, R. (2026). Online Examination Security: A Multi-Layered Threat Framework and AI-Based Proctoring Architecture for Academic Integrity. Dandao Xuebao/Journal of Ballistics, 38(1), 418-427. https://doi.org/10.52783/dxjb.v38.291

Vancouver

Trivedi R. Online Examination Security: A Multi-Layered Threat Framework and AI-Based Proctoring Architecture for Academic Integrity. Dandao Xuebao/Journal of Ballistics. 2026;38(1):418-427. doi:10.52783/dxjb.v38.291

BibTeX
@article{Trivedi2026,
  author = {Trivedi, Rushit},
  title = {Online Examination Security: A Multi-Layered Threat Framework and AI-Based Proctoring Architecture for Academic Integrity},
  journal = {Dandao Xuebao/Journal of Ballistics},
  year = {2026},
  volume = {38},
  number = {1},
  pages = {418--427},
  doi = {10.52783/dxjb.v38.291}
}
Kostenlos starten →