Can We Trust Large Language Models Generated Code? A Framework for In-Context Learning, Security Patterns, and Code Evaluations Across Diverse LLMs

TL;DR

The paper examines the security of code generated by large language models (LLMs) and proposes a framework that employs In-Context Learning (ICL) to imbue security knowledge into code generation. It compares four LLM platforms (two prompt-driven and two coding copilots) across three languages and uses SAST, code reviews, and a Code Security Risk Measure (CSRM) to quantify vulnerabilities and code smells. Findings show that one-shot and few-shot ICL can reduce security weaknesses, though substantial vulnerabilities remain and some models exhibit learning biases or introduce supply-chain risks. The work highlights the need for ongoing security-focused evaluation, human-in-the-loop review, and model training strategies to mitigate new attack vectors arising from AI-assisted code generation. It also provides datasets and methodological groundwork for future research on secure LLM-based code generation and repair.

Abstract

Large Language Models (LLMs) such as ChatGPT and GitHub Copilot have revolutionized automated code generation in software engineering. However, as these models are increasingly utilized for software development, concerns have arisen regarding the security and quality of the generated code. These concerns stem from LLMs being primarily trained on publicly available code repositories and internet-based textual data, which may contain insecure code. This presents a significant risk of perpetuating vulnerabilities in the generated code, creating potential attack vectors for exploitation by malicious actors. Our research aims to tackle these issues by introducing a framework for secure behavioral learning of LLMs through In-Content Learning (ICL) patterns during the code generation process, followed by rigorous security evaluations. To achieve this, we have selected four diverse LLMs for experimentation. We have evaluated these coding LLMs across three programming languages and identified security vulnerabilities and code smells. The code is generated through ICL with curated problem sets and undergoes rigorous security testing to evaluate the overall quality and trustworthiness of the generated code. Our research indicates that ICL-driven one-shot and few-shot learning patterns can enhance code security, reducing vulnerabilities in various programming scenarios. Developers and researchers should know that LLMs have a limited understanding of security principles. This may lead to security breaches when the generated code is deployed in production systems. Our research highlights LLMs are a potential source of new vulnerabilities to the software supply chain. It is important to consider this when using LLMs for code generation. This research article offers insights into improving LLM security and encourages proactive use of LLMs for code generation to ensure software system safety.

Figures (13)

• Figure 1: The LLM code generation process: simplified Transformer architecture for code generation with potential security risks
• Figure 2: Prompt Driven Code Generators with BLLMs
• Figure 3: Coding CoPilots LLMs for Code Generation
• Figure 4: Research Questions
• Figure 5: Large Language Models Code Evaluations Framework using ICL Security Patterns