EduZone: A Framework for Evaluating LLM Safety for K-12 Students and Teachers

2026-08-03

AI Summary

This study addresses the critical gap in evaluating the safety of large language models (LLMs) within K–12 educational settings, where student–teacher interactions may elicit harmful content that existing assessments overlook. To this end, the authors propose EduZone, a novel framework that systematically integrates educational context, curriculum-aligned knowledge points, and 28 education-specific risk criteria across six categories to construct both single-turn and dynamic multi-turn adversarial interactions. EduZone further introduces a four-tiered safety response evaluation protocol. Experimental results demonstrate that mainstream LLMs are significantly more prone to failure when confronted with education-related risks and evolving multi-turn dialogues, revealing inadequacies in current safety mechanisms. EduZone thus provides a scalable, automated benchmark for assessing LLM safety in educational contexts, effectively filling a crucial void in domain-specific evaluation methodologies.

Abstract

Large language models (LLMs) are increasingly used across diverse tasks in K-12 education, yet existing safety evaluations rarely examine how harmful or inappropriate content appears in interactions between LLMs and students or teachers. To address this, we present EduZone, an evaluation framework for LLM safety across diverse educational scenarios. Our framework systematically combines (1) student- and teacher-facing LLM usage contexts, (2) fine-grained curriculum concepts, and (3) 6 risk categories and 28 subcategories spanning both conventional and education-specific harms to generate contextually grounded adversarial interactions. We construct these interactions in three settings: single-turn requests, static multi-turn conversations, and dynamic multi-turn conversations. Using these interactions, we evaluate ten LLMs using four safety levels: refusal, safe assistance, risky assistance with safety guidance, and fully risky assistance. Our results reveal greater vulnerability to education-specific risks and dynamic multi-turn interactions, while existing safety guardrails fail to adequately address these risks. EduZone advances LLM safety in education by providing an automated, scalable evaluation framework that supports the development and deployment of safer LLMs in K-12 education.

Problem

• LLM safety
• K-12 education
• adversarial interactions
• education-specific risks
• safety evaluation

Innovation

• LLM safety evaluation
• education-specific risks
• adversarial interactions
• multi-turn conversations
• K-12 education