Generative AI halves coding time but floods open source with low-quality pull requests

Generative artificial intelligence tools are significantly reducing the time developers spend on coding tasks, yet this efficiency comes with a notable downside for open-source communities. According to a report by Neowin, the rapid generation of code is flooding repositories with low-quality contributions, creating new maintenance burdens for project maintainers who must sift through these submissions.

The data indicates that while the volume of contributions is rising, the quality is not keeping pace. Neowin reports that twenty-five percent of incoming pull requests are deemed low-quality slop. This influx of subpar code requires maintainers to invest additional time in review and rejection processes, potentially slowing down the integration of valuable, high-quality contributions from human developers.

Impact on Development Efficiency

The impact of generative AI on development speed varies depending on the complexity of the task. For many routine coding activities, the technology has proven effective, cutting coding time in half for a significant portion of users. This acceleration allows teams to iterate faster and potentially deliver features more quickly, provided the generated code meets the project's standards and requires minimal refactoring.

However, the benefits are not universal across all types of development work. Neowin notes that for twenty-one percent of respondents, high-complexity tasks actually took longer when using AI tools. This suggests that while generative models excel at straightforward, repetitive coding patterns, they may struggle with nuanced, complex logic, requiring developers to spend more time correcting errors or refining the output than they would have spent writing the code manually.

Maintainer Workload and Contribution Trends

Despite the quality concerns, the overall volume of engagement with open-source projects appears to be increasing. According to Neowin, fifty-nine percent of maintainers report growing contributions to their projects. This trend indicates that developers are more active than ever, likely aided by the lower barrier to entry provided by AI-assisted coding tools. The challenge for maintainers is balancing this increased activity with the need to preserve code quality and project integrity.

The term "slop" refers to the low-quality, often repetitive or contextually inappropriate code generated by AI models. This content can clutter issue trackers and pull request queues, making it harder for maintainers to identify genuine improvements or bug fixes. The rise of AI-generated slop highlights the growing need for better automated filtering tools and more robust community guidelines to manage the influx of machine-generated contributions.

What should you do?

Project maintainers and security professionals should monitor their repositories for signs of AI-generated low-quality code. Implementing stricter code review processes and utilizing automated static analysis tools can help identify and filter out slop before it enters the codebase. Developers should also verify AI-generated code thoroughly, especially for complex tasks, to ensure it meets security and performance standards.

Background: Generative AI

Generative AI examples include text completion, image synthesis and code generation. These systems predict the next likely token based on vast training datasets. Understanding the underlying mechanics helps you identify risks like data leakage and hallucination in your infrastructure.

Read the full guide: Generative AI Examples: Mechanisms, Risks and Control

What to do and how to stay safe: generative AI

• Implement automated static analysis tools to detect low-quality or suspicious code patterns in pull requests.
• Establish clear community guidelines regarding the use of AI-generated code in contributions.
• Require manual review for all pull requests involving complex logic or security-sensitive components.
• Monitor contribution metrics to identify sudden spikes in low-quality submissions that may indicate AI abuse.

Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation

General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

How much of incoming pull requests are low-quality?

According to Neowin, twenty-five percent of incoming pull requests are deemed low-quality slop.

Does AI always speed up coding?

No, twenty-one percent of respondents found that high-complexity tasks took longer when using AI tools.

Are contributions to open source increasing?

Yes, fifty-nine percent of maintainers report growing contributions to their projects.