# AI Code in the Wild: Measuring Security Risks and Ecosystem Shifts of AI-Generated Code in Modern Software

arXiv: https://arxiv.org/abs/2512.18567 · HTML: https://mlsanity.com/paper/2512.18567/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 34 of its 59 references that are in the map.
- First posted on arXiv 21 Dec 2025.

## Builds on (most cited first, 30 of 34)

- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [CodeBERT: A Pre-Trained Model for Programming and Natural Languages](https://mlsanity.com/paper/2002.08155/) (2020)
- [The Impact of AI on Developer Productivity: Evidence from GitHub Copilot](https://mlsanity.com/paper/2302.06590/) (2023)
- [Do Users Write More Insecure Code with AI Assistants?](https://mlsanity.com/paper/2211.03622/) (2022)
- [A Survey on Code Generation with LLM-based Agents](https://mlsanity.com/paper/2508.00083/) (2025)
- [Natural Attack for Pre-trained Models of Code](https://mlsanity.com/paper/2201.08698/) (2022)
- [Evaluating the Code Quality of AI-Assisted Code Generation Tools: An Empirical Study on GitHub Copilot, Amazon CodeWhisperer, and ChatGPT](https://mlsanity.com/paper/2304.10778/) (2023)
- [CodeLMSec Benchmark: Systematically Evaluating and Finding Security Vulnerabilities in Black-Box Code Language Models](https://mlsanity.com/paper/2302.04012/) (2023)
- [Security Weaknesses of Copilot-Generated Code in GitHub Projects: An Empirical Study](https://mlsanity.com/paper/2310.02059/) (2023)
- [Automatic Programming: Large Language Models and Beyond](https://mlsanity.com/paper/2405.02213/) (2024)
- [Vulnerabilities in AI Code Generators: Exploring Targeted Data Poisoning Attacks](https://mlsanity.com/paper/2308.04451/) (2023)
- [CodeIP: A Grammar-Guided Multi-Bit Watermark for Large Language Models of Code](https://mlsanity.com/paper/2404.15639/) (2024)
- [SecVulEval: Benchmarking LLMs for Real-World C/C++ Vulnerability Detection](https://mlsanity.com/paper/2505.19828/) (2025)
- [Assessing AI Detectors in Identifying AI-Generated Code: Implications for Education](https://mlsanity.com/paper/2401.03676/) (2024)
- [Transforming Software Development: Evaluating the Efficiency and Challenges of GitHub Copilot in Real-World Projects](https://mlsanity.com/paper/2406.17910/) (2024)
- [Exploring the Limitations of Detecting Machine-Generated Text](https://mlsanity.com/paper/2406.11073/) (2024)
- [A.S.E: A Repository-Level Benchmark for Evaluating Security in AI-Generated Code](https://mlsanity.com/paper/2508.18106/) (2025)
- [MCPGuard : Automatically Detecting Vulnerabilities in MCP Servers](https://mlsanity.com/paper/2510.23673/) (2025)
- [Just another copy and paste? Comparing the security vulnerabilities of ChatGPT generated code and StackOverflow answers](https://mlsanity.com/paper/2403.15600/) (2024)
- [The Impact of Generative AI on Collaborative Open-Source Software Development: Evidence from GitHub Copilot](https://mlsanity.com/paper/2410.02091/) (2024)
- [LLMs: A Game-Changer for Software Engineers?](https://mlsanity.com/paper/2411.00932/) (2024)
- [Security Degradation in Iterative AI Code Generation -- A Systematic Analysis of the Paradox](https://mlsanity.com/paper/2506.11022/) (2025)
- [textttDroid: A Resource Suite for AI-Generated Code Detection](https://mlsanity.com/paper/2507.10583/) (2025)
- [Assessing the Quality and Security of AI-Generated Code: A Quantitative Analysis](https://mlsanity.com/paper/2508.14727/) (2025)
- [Predicting the Popularity of GitHub Repositories](https://mlsanity.com/paper/1607.04342/) (2016)
- [Enhancing Security of AI-Based Code Synthesis with GitHub Copilot via Cheap and Efficient Prompt-Engineering](https://mlsanity.com/paper/2403.12671/) (2024)
- [ChatGPT Code Detection: Techniques for Uncovering the Source of Code](https://mlsanity.com/paper/2405.15512/) (2024)
- [CodeMirage: A Multi-Lingual Benchmark for Detecting AI-Generated and Paraphrased Source Code from Production-Level LLMs](https://mlsanity.com/paper/2506.11059/) (2025)
- [RefleXGen:The unexamined code is not worth using](https://mlsanity.com/paper/2510.23674/) (2025)
- [Is Your Prompt Poisoning Code? Defect Induction Rates and Security Mitigation Strategies](https://mlsanity.com/paper/2510.22944/) (2025)

## Built on by (most cited first, 4 of 4)

- [Correct Code, Vulnerable Dependencies: A Large Scale Measurement Study of LLM-Specified Library Versions](https://mlsanity.com/paper/2605.06279/) (2026)
- [The Substrate Collapse: AI Code Generation Invalidates Authorship-Based Knowledge Metrics](https://mlsanity.com/paper/2606.20882/) (2026)
- [Activation Probes Surface Code-Security Signals that the Model's Output Misses](https://mlsanity.com/paper/2608.09643/) (2026)
- [Generation of Web Apps with Agentic IDEs: An Empirical Assessment](https://mlsanity.com/paper/2608.20903/) (2026)

## Abstract (arXiv)

Large language models (LLMs) for code generation are becoming integral to modern software development, but their real-world prevalence and security impact remain poorly understood. We present the first large-scale empirical study of AI-generated code (AIGCode) in the wild. We build a high-precision detection pipeline and a representative benchmark to distinguish AIGCode from human-written code, and apply them to (i) development commits from the top 1,000 GitHub repositories (2022-2025) and (ii) 7,000+ recent CVE-linked code changes. This lets us label commits, files, and functions along a human/AI axis and trace how AIGCode moves through projects and vulnerability life cycles. Our measurements show three ecological patterns. First, AIGCode is already a substantial fraction of new code, but adoption is structured: AI concentrates in glue code, tests, refactoring, documentation, and other boilerplate, while core logic and security-critical configurations remain mostly human-written. Second, adoption has security consequences: some CWE families are overrepresented in AI-tagged code, and near-identical insecure templates recur across unrelated projects, suggesting "AI-induced vulnerabilities" propagated by shared models rather than shared maintainers. Third, in human-AI edit chains, AI introduces high-throughput changes while humans act as security gatekeepers; when review is shallow, AI-introduced defects persist longer, remain exposed on network-accessible surfaces, and spread to more files and repositories. We will open-source the complete dataset and release analysis artifacts and fine-grained documentation of our methodology and findings.