| Exploring the Potential and Obstacles of Large Language Models in Enhancing Code Security |
| کد مقاله : 1074-NAEC |
| نویسندگان |
|
شیدا دوه لی * گروه مهندسی کامپیوتر، دانشکده ی فنی و مهندسی، دانشگاه بین المللی اهل بیت، تهران، ایران. |
| چکیده مقاله |
| Large Language Models (LLMs) have ushered in a transformative era in software engineering, particularly in enhancing code security, which is increasingly crucial in today’s digital landscape. This systematic literature review (SLR) consolidates a comprehensive array of recent findings to delve into the myriad potential benefits and challenges associated with the application of LLMs in improving software security practices. Through a meticulous analysis of 35 studies published between 2020 and March 2025, we identify significant advantages of LLMs, including their ability to detect vulnerabilities more accurately, automate code evaluations, and facilitate the generation of secure code. However, the review also addresses critical challenges that accompany their implementation, such as inaccuracies in model outputs, inherent biases present in training datasets, considerable computational resource requirements, and pressing ethical dilemmas. By synthesizing these insights, this review serves as a vital resource for researchers and software developers aiming to harness LLMs effectively while navigating the complexities of maintaining robust security measures in software development. |
| کلیدواژه ها |
| Software Engineering, Large Language Model, Ethical Artificial Intelligence, Secure Code. |
| وضعیت: پذیرفته شده |
