DEVELOPING A SECURE AI-BASED EXAMINATION MANAGEMENT PLATFORM FOR SECONDARY SPECIAL EDUCATION

Objective: To design, implement, and empirically evaluate a secure, accessible, AI-driven examination management platform specifically engineered for secondary special education environments, addressing accessibility barriers, personalized item adaptation, and assessment security. Material and Methods: The platform architecture integrates four core modules: an AI Engine for adaptive testing and automated grading, a Security Layer featuring multi-factor authentication, AES-256 encryption, and AI-driven anti-cheating protocols, a Database Layer for encrypted storage, and an accessible User Interface complying with WCAG 2.1 AA standards. The platform was evaluated across secondary special education institutions focusing on security, efficiency, accessibility, and accuracy. Results: Empirical evaluations demonstrated outstanding performance across key metrics: 95% overall security (95%+ reduction in cheating attempts with zero data breaches), 90% usability and accessibility satisfaction (WCAG 2.1 AA compliance), 92% performance efficiency (<2s response time, 99.9% uptime, supporting 10,000+ concurrent users), and 94% system reliability. Compared to traditional examination systems, the proposed platform achieved a +35% increase in security, +35% in operational efficiency, +47% in accessibility, and +24% in assessment accuracy. Conclusion: The developed platform successfully resolves the operational dilemma between rigorous exam security and inclusive accessibility in secondary special education, establishing a scalable, highly reliable framework for modern digital assessment.