一句话：梳理 AI 与大语言模型在软件开发生命周期各阶段的应用、风险与挑战，并提出未来研究议程。
随着人工智能与大语言模型（LLM）的快速发展，代码生成、调试、测试与维护等软件工程任务正被逐步自动化，ChatGPT、DeepSeek 等 AI 工具显示出提升开发者生产力、加速开发流程的潜力。然而，将 AI 与 LLM 引入软件工程也带来显著风险：AI 生成代码的可靠性难以保证、可能引入安全漏洞、训练数据中的偏见会被传播；同时知识产权、透明性与人工监督等伦理问题增加了在关键系统中采用 AI 的复杂度。本文分析 AI 在软件工程中的角色，考察其在软件开发生命周期各阶段的影响，识别关键应用、挑战与未来研究方向，为 AI 驱动的软件工程研究议程提供参考。
The rapid advancements in Artificial Intelligence (AI) and Large Language Models (LLMs) are reshaping software engineering and automating tasks such as code generation, debugging, testing, and maintenance. AI-powered tools (i.e. ChatGPT, DeepSeek), have demonstrated significant potential in enhancing developer productivity and accelerating software development processes. Integrating AI and LLMs into software engineering presents notable challenges despite these advancements. Concerns regarding the reliability of AI-generated code, security vulnerabilities, and the propagation of biases in training data pose substantial risks. Additionally, ethical considerations, including intellectual property rights, transparency, and the need for human oversight, highlight the complexities of AI adoption in critical software systems. The rapid evolution of these technologies requires continuous adaptation of software engineering methodologies to mitigate risks while maximizing benefits. This paper analyzes AI's role in software engineering, identifying key applications, challenges, and future research directions. We examine AI's impact across various phases of the software development lifecycle, This paper contributes to the ongoing discussion on AI-driven software engineering and outlines a research agenda for navigating this rapidly evolving field.
DOI 原文 ·
@inproceedings{paperbot8954,
title = {AI in the Software Development Lifecycle: Insights and Open Research Questions},
author = {Everton Guimaraes and Nathalia Nascimento},
booktitle = {FSE},
year = {2025},
doi = {10.1145/3696630.3730538}
}