Artificial Intelligence and the Future of Assessment: Opportunities for Scalable, Fair, and Real-Time Evaluation

الملخص

Artificial intelligence (AI) is rapidly transforming educational assessment by enabling automated grading, adaptive testing, and personalized feedback at unprecedented scale. Research suggests that AI tools can enhance assessment scalability by automating routine grading tasks and supporting large-scale testing programs, while also enabling real-time feedback to learners (e.g., Shute, 2008; U.S. Dept. of Ed., 2023). At the same time, AI promises to improve fairness by reducing human grader bias and standardizing scoring across diverse student populations. Early deployments range from computer‐scored language proficiency exams to AI-driven peer review systems in MOOCs and adaptive formative platforms. Case studies in China, the United States, and international programs demonstrate these opportunities as well as significant challenges. On the one hand, studies find high agreement between AI and human scores in some contexts (e.g., ~92% agreement in a Chinese AI grading system) and large time savings (Gradescope claims 90% faster grading). On the other hand, rigorous evaluations (e.g., UNESCO and independent studies) have identified fairness, validity, and transparency issues. This paper reviews the theoretical foundations of AI-enabled assessment and surveys global examples of AI use in evaluation. We analyze how AI can provide scalable, equitable, and real-time formative and summative assessment, and we discuss practical limitations, ethical risks (including algorithmic bias, data privacy, and digital divides), and implementation challenges. Finally, we offer recommendations for policy and practice to guide responsible, effective integration of AI into education systems.