Imagen de Portada

Position: The Pitfalls of Over-Alignment: Overly Caution Health-Related Responses From LLMs are Unethical and Dangerous

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Guo, Wenqi Marshall, Du, Yiyang, Tworek, Heidi J. S., Du, Shan
Formato:	Preprint
Publicado:	2025
Materias:	Computers and Society
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Position: Universal Aesthetic Alignment Narrows Artistic Expression
por: Guo, Wenqi Marshall, et al.
Publicado: (2025)

LangGas: Introducing Language in Selective Zero-Shot Background Subtraction for Semi-Transparent Gas Leak Detection with a New Dataset
por: Guo, Wenqi, et al.
Publicado: (2025)

Pluralistic Alignment Over Time
por: Klassen, Toryn Q., et al.
Publicado: (2024)

Mind the Gap: Pitfalls of LLM Alignment with Asian Public Opinion
por: Shankar, Hari, et al.
Publicado: (2026)

Evaluating the Promise and Pitfalls of LLMs in Hiring Decisions
por: Anzenberg, Eitan, et al.
Publicado: (2025)

AI Identity, Empowerment, and Mindfulness in Mitigating Unethical AI Use
por: Shaayesteh, Mayssam Tarighi, et al.
Publicado: (2025)

Take Caution in Using LLMs as Human Surrogates: Scylla Ex Machina
por: Gao, Yuan, et al.
Publicado: (2024)

Safe in the Future, Dangerous in the Past: Dissecting Temporal and Linguistic Vulnerabilities in LLMs
por: Said, Muhammad Abdullahi, et al.
Publicado: (2025)

Navigating Pitfalls: Evaluating LLMs in Machine Learning Programming Education
por: Kumar, Smitha, et al.
Publicado: (2025)

The AI Risk Spectrum: From Dangerous Capabilities to Existential Threats
por: Grey, Markov, et al.
Publicado: (2025)

TikTok Engagement Traces Over Time and Health Risky Behaviors: Combining Data Linkage and Computational Methods
por: Zhao, Xinyan, et al.
Publicado: (2024)

Toward Responsible and Beneficial AI: Comparing Regulatory and Guidance-Based Approaches -A Comprehensive Comparative Analysis of Artificial Intelligence Governance Frameworks across the European Union, United States, China, and IEEE
por: Du, Jian
Publicado: (2025)

No Size Fits All: The Perils and Pitfalls of Leveraging LLMs Vary with Company Size
por: Urlana, Ashok, et al.
Publicado: (2024)

Dynamic Bayesian Item Response Model with Decomposition (D-BIRD): Modeling Cohort and Individual Learning Over Time
por: Lee, Hansol, et al.
Publicado: (2025)

FaceLinkGen: Rethinking Identity Leakage in Privacy-Preserving Face Recognition with Identity Extraction
por: Guo, Wenqi, et al.
Publicado: (2026)

Exploring Safety Alignment Evaluation of LLMs in Chinese Mental Health Dialogues via LLM-as-Judge
por: Cai, Yunna, et al.
Publicado: (2025)

Pitfalls of Evidence-Based AI Policy
por: Casper, Stephen, et al.
Publicado: (2025)

VSF: Simple, Efficient, and Effective Negative Guidance in Few-Step Image Generation Models By Value Sign Flip
por: Guo, Wenqi, et al.
Publicado: (2025)

Boosting Fairness and Robustness in Over-the-Air Federated Learning
por: Oksuz, Halil Yigit, et al.
Publicado: (2024)

Faults and Pitfalls in Implementing the Right to be Forgotten
por: Sun, Chen, et al.
Publicado: (2026)

Large Language Models as Students Who Think Aloud: Overly Coherent, Verbose, and Confident
por: Borchers, Conrad, et al.
Publicado: (2026)

Verify with Caution: The Pitfalls of Relying on Imperfect Factuality Metrics
por: Godbole, Ameya, et al.
Publicado: (2025)

Personalized Parsons Puzzles as Scaffolding Enhance Practice Engagement Over Just Showing LLM-Powered Solutions
por: Hou, Xinying, et al.
Publicado: (2025)

The Disintegration of Free Speech
por: Mei, Yiyang
Publicado: (2026)

Risks of AI Scientists: Prioritizing Safeguarding Over Autonomy
por: Tang, Xiangru, et al.
Publicado: (2024)

Impact of AI Tools on Learning Outcomes: Decreasing Knowledge and Over-Reliance
por: Benedek, Márton, et al.
Publicado: (2025)

Access Over Deception: Fighting Deceptive Patterns through Accessibility
por: Pellkvist, Tobias, et al.
Publicado: (2026)

Out of the Loop Again: How Dangerous is Weaponizing Automated Nuclear Systems?
por: Schwartz, Joshua A., et al.
Publicado: (2025)

Expanding External Access To Frontier AI Models For Dangerous Capability Evaluations
por: Charnock, Jacob, et al.
Publicado: (2026)

Technical Requirements for Halting Dangerous AI Activities
por: Barnett, Peter, et al.
Publicado: (2025)

Recommendation Fairness in Social Networks Over Time
por: Cao, Meng, et al.
Publicado: (2024)

Fluent but Foreign: Even Regional LLMs Lack Cultural Alignment
por: Agarwal, Dhruv, et al.
Publicado: (2025)

Randomness, Not Representation: The Unreliability of Evaluating Cultural Alignment in LLMs
por: Khan, Ariba, et al.
Publicado: (2025)

Measuring Compliance with the California Consumer Privacy Act Over Space and Time
por: Tran, Van, et al.
Publicado: (2024)

Reclaiming Constitutional Authority of Algorithmic Power
por: Mei, Yiyang, et al.
Publicado: (2025)

An FDA for AI? Pitfalls and Plausibility of Approval Regulation for Frontier Artificial Intelligence
por: Carpenter, Daniel, et al.
Publicado: (2024)

Optimizing Mastery Learning by Fast-Forwarding Over-Practice Steps
por: Xia, Meng, et al.
Publicado: (2025)

Evaluating the Clinical Safety of LLMs in Response to High-Risk Mental Health Disclosures
por: Shah, Siddharth, et al.
Publicado: (2025)

Robustness and Confounders in the Demographic Alignment of LLMs with Human Perceptions of Offensiveness
por: Alipour, Shayan, et al.
Publicado: (2024)

Understanding Cultural Alignment in Multilingual LLMs via Natural Debate Statements
por: Negru, Vlad-Andrei, et al.
Publicado: (2026)