Date
AI Overview
Why it matters
Teachers face high ethical demands that affect education quality, but comprehensive psychological and ethical evaluations of large language models simulating teacher roles are scarce.
Study at a glance
- Question: How do large language models simulating teachers align with human educator personalities, moral reasoning, and safety risks?
- Setting and data: 100 in-service teachers, 9 expert annotators, and 14 large language models evaluated on moral dilemmas and prompt injections.
- Method: Model evaluation benchmark using personality scales, expert ratings of moral dilemmas, and adversarial prompt tests.
- AI tools: GPT-4.1, DeepSeek-R1, Claude-3.7, Qwen3-235B-R, GLM-Z1-9B, Baichuan2-7B
Key findings
- Simulated teacher models showed more idealized and polarized personalities than real teachers, scoring higher on Honesty-Humility but lower on Emotionality.
- Reasoning-oriented models demonstrated significantly higher moral reasoning stages than non-reasoning models, particularly on abstract value conflicts.
- Stronger reasoning models were more vulnerable to harmful soft prompt injections, revealing a capability-safety paradox confirmed in controlled testing.
Summary drafted with AI from the full paper and may contain errors. See the full paper for details.
Simulating Professions (SP) enables Large Language Models (LLMs) to emulate professional roles. However, comprehensive psychological and ethical evaluation in these contexts remains lacking. This paper introduces EMNLP, an Educator-role Moral and Normative LLMs Profiling framework for personality profiling, moral development stage measurement, and ethical risk under soft prompt injection. EMNLP extends existing scales and constructs 88 teacher-specific moral dilemmas, enabling profession-oriented comparison with human teachers. A targeted soft prompt injection set evaluates compliance and vulnerability in teacher SP. Experiments on 14 LLMs show teacher-role LLMs exhibit more idealized and polarized personalities than human teachers, excel in abstract moral reasoning, but struggle with emotionally complex situations. Models with stronger reasoning are more vulnerable to harmful prompt injection, revealing a paradox between capability and safety. The model temperature and other hyperparameters have limited influence except in some risk behaviors. This paper presents the first benchmark to assess ethical and psychological alignment of teacher-role LLMs for educational AI. Resources are available at https://e-m-n-l-p.github.io/.
Research Categories
What is the application?
Who is the user?
Why use AI?
Study design
Methods
AI tools
Publication Source
Publisher
arXiv
