Beyond Context: Large Language Models' Failure to Grasp Users' Intent
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910163461472256 |
|---|---|
| author | Hussain, Ahmed M. Salahuddin, Salahuddin |
| author_facet | Hussain, Ahmed M. Salahuddin, Salahuddin |
| contents | Current Large Language Models (LLMs) safety approaches focus on explicitly harmful content while overlooking a critical vulnerability: the inability to understand context and recognize user intent. This creates exploitable vulnerabilities that malicious users can systematically leverage to circumvent safety mechanisms. We empirically evaluate multiple state-of-the-art LLMs, including ChatGPT, Claude, Gemini, and DeepSeek. Our analysis demonstrates the circumvention of reliable safety mechanisms through emotional framing, progressive revelation, and academic justification techniques. Notably, reasoning-enabled configurations amplified rather than mitigated the effectiveness of exploitation, increasing factual precision while failing to interrogate the underlying intent. The exception was Claude Opus 4.1, which prioritized intent detection over information provision in some use cases. This pattern reveals that current architectural designs create systematic vulnerabilities. These limitations require paradigmatic shifts toward contextual understanding and intent recognition as core safety capabilities rather than post-hoc protective mechanisms. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_21110 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Beyond Context: Large Language Models' Failure to Grasp Users' Intent Hussain, Ahmed M. Salahuddin, Salahuddin Artificial Intelligence Computation and Language Cryptography and Security Computers and Society Current Large Language Models (LLMs) safety approaches focus on explicitly harmful content while overlooking a critical vulnerability: the inability to understand context and recognize user intent. This creates exploitable vulnerabilities that malicious users can systematically leverage to circumvent safety mechanisms. We empirically evaluate multiple state-of-the-art LLMs, including ChatGPT, Claude, Gemini, and DeepSeek. Our analysis demonstrates the circumvention of reliable safety mechanisms through emotional framing, progressive revelation, and academic justification techniques. Notably, reasoning-enabled configurations amplified rather than mitigated the effectiveness of exploitation, increasing factual precision while failing to interrogate the underlying intent. The exception was Claude Opus 4.1, which prioritized intent detection over information provision in some use cases. This pattern reveals that current architectural designs create systematic vulnerabilities. These limitations require paradigmatic shifts toward contextual understanding and intent recognition as core safety capabilities rather than post-hoc protective mechanisms. |
| title | Beyond Context: Large Language Models' Failure to Grasp Users' Intent |
| topic | Artificial Intelligence Computation and Language Cryptography and Security Computers and Society |
| url | https://arxiv.org/abs/2512.21110 |