Evaluating Prompt Injection Defenses for Educational LLM Tutors: Security-Usability-Latency Trade-offs

Evaluating Prompt Injection Defenses for Educational LLM Tutors: Security-Usability-Latency Trade-offs Maiorano (2026) — arXiv cs.CR/cs.AI.

Evaluating Prompt Injection Defenses for Educational LLM Tutors: Security-Usability-Latency Trade-offs

Summary

This paper addresses a fundamental tension in educational LLM deployment: tutors must be both secure (resist prompt injection attacks) and usable (not block legitimate educational interactions). The author presents a systematic evaluation methodology using a 480-query benchmark (369 injection, 111 benign) with statistically rigorous comparison.

Defense methods compared:

The proposed pipeline combines deterministic pattern filters, structural validation, contextual sandboxing, and session-level behavioral checks. Its design prioritizes pedagogical usability — zero false positives means no legitimate student queries get blocked, an essential requirement for Intelligent Tutoring systems where interruptions harm learning.

NeMo Guardrails blocks all attacks but incorrectly flags ~16% of benign requests — a rate that would seriously degrade the Student Experience in real tutoring sessions. Prompt Guard provides middle-ground performance.

The framework enables evidence-based guardrail selection under institutional risk and usability requirements. This directly connects to AI Tutor Safety Harms concerns and the emerging field of Pedagogical Safety RL. The latency dimension is particularly important for real-time Conversational AI Tutors Framework where response delays degrade engagement.

The paper highlights that educational settings have unique requirements: false positives are more costly than in general-purpose chatbots, because blocking a student's learning interaction carries pedagogical harm. This aligns with findings in Eduframetrap LLM Sycophancy Educational Safety that educational safety requires domain-specific benchmarks.

Connected Concepts

Connected Articles

Citation

Maiorano, A. C. (2026). Evaluating prompt injection defenses for educational LLM tutors: Security-usability-latency trade-offs. arXiv:2605.06669.