Initially, many were skeptical of artificial intelligence, especially when image and video creations resembled memes. Some doubted AI’s potential for improvement, viewing it as a waste of time and money. However, AI is now increasingly used and even depended upon, which is not a cause for celebration. In fact, Linux kernel developers are tired of receiving AI-generated bug reports because they contain errors, inconsistencies, and are difficult to understand.
The primary advantage of AI was to increase productivity by offloading repetitive and tedious tasks, allowing humans to focus on other work. At least, that was the theory. The practice, however, falls far short, as AI still requires human supervision, and even AI agents promising the automation businesses desire end up making mistakes.
Linux Developers are Saturated with AI-Generated Bug Reports, Which Contain Errors and Take More Time to Understand and Fix
With the rise of AI, more and more people are using it, and the problem is the emergence of dependency. This means some individuals need AI to do their work or even for daily tasks, and without this assistant, they find everything much harder or are incapable of performing their duties. This has been observed in various sectors, such as programming, where AI is overused, promoting “vibe coding.” Linux kernel developers themselves are experiencing this, but from the other side, as they have to review hundreds and thousands of AI-generated bug reports.
According to them, this is happening because users rely on AI models and, of course, ask them to generate bug and vulnerability reports, which are then submitted as-is for review. The issue is that AI makes mistakes, and ironically, these bug reports contain further errors, with nonsensical or mismatched code lines and explanations. Developers bear the brunt of this, as reviewing such reports ends up taking far longer than it would for a human-created report.
The Open Source Community Experiences AI as a Double-Edged Sword, Helping Identify More Errors When Used Correctly, but Slowing Down the Process Otherwise
This situation has affected not only Linux developers but also developers of open-source projects and software. While the idea of an Open Source community is that anyone can contribute and/or help a project, using AI to generate reports turns it into a double-edged sword. On one hand, with the help of AI, many more errors and vulnerabilities can be discovered, especially when using cybersecurity-focused models like Claude Mythos.
On the other hand, AI is generating sub-optimal bug reports. These not only slow down the review process but also generate reports at a much faster rate than a person could manually write them. Consequently, they are overwhelmed with low-quality and unreliable documents. In turn, developers could use AI to speed up the verification process, but this would lead back to the problem that AI is not perfect and makes mistakes.
