August 12, 2025

Security of Language Models for Code: A Systematic Literature Review

Key Points

MAIN FINDING: Language models for code show significant promise but are vulnerable to security risks.
KEY EVIDENCE: The review includes 68 studies organized by attack and defense strategies in language models.
APPROACH: Comprehensive survey of literature covering models, datasets, and evaluation metrics in code security.
SIGNIFICANCE: This work fills a crucial gap, offering insights for enhancing cybersecurity in software engineering.

Abstract

Language models for code (CodeLMs) have emerged as powerful tools for code-related tasks, outperforming traditional methods and standard machine learning approaches. However, these models are susceptible to security vulnerabilities, drawing increasing research attention from domains such as software engineering, artificial intelligence, and cybersecurity. Despite the growing body of research focused on the security of CodeLMs, a comprehensive survey in this area remains absent. To address this gap, we systematically review 68 relevant papers, organizing them based on attack and defense strategies. Furthermore, we provide an overview of commonly used language models, datasets, and evaluation metrics, and highlight open-source tools and promising directions for future research in securing CodeLMs.

Perguntar à IA

Bookmark