The development team behind the compiler 'GCC' has published its acceptance policy for AI-generated code, rejecting AI-generated code longer than 15 lines.
While AI-generated code is becoming more common, open-source projects are facing the problem of maintainers becoming overwhelmed by a massive influx of AI-generated pull requests. The development team of the open-source compiler 'GNU Compiler Collection (GCC)' has now published its acceptance policy for AI-generated code.
GNU Compiler Collection - AI Policy - GNU Project
GCC Compiler Bans AI Code Contribution But Sensibly
https://itsfoss.com/news/gcc-bans-ai-code/
The GCC's AI policy was adopted on July 29, 2026. The AI policy explicitly states that it rejects all 'legally significant contributions including content generated under an LLM' or 'legally significant contributions derived from content generated under an LLM.'
According to the GNU maintainer guidelines, a 'legally significant contribution' is defined as 'a contribution of more than approximately 15 lines of code or text,' and contributions exceeding 15 lines are subject to copyright. In other words, GCC has adopted a policy of rejecting AI-generated content exceeding 15 lines.
On the other hand, code and text that do not meet the definition of a 'legally significant contribution' are acceptable. However, AI-generated output must be tagged with 'Assisted-by:' to clearly indicate that AI was used. Additionally, AI-generated output is accepted as an exception in test cases.
The mechanism of adding 'Assisted-by:' to AI output is also used in the Linux kernel development project.
The Linux kernel will now place legal responsibility on the person who submits the application for 'all lines of code generated by AI, and any resulting bugs or security flaws' - GIGAZINE
by Mike Cohen
In the case of the Linux kernel development project, they have introduced an AI-powered security scanning system called ' Sashiko ,' demonstrating a proactive approach to incorporating AI into development. While some project participants believe that AI should not be used even for security scanning, Linus Torvalds has shown a willingness to adopt AI as a highly convenient technology.
Linus Torvalds gives his opinion on the use of AI in Linux kernel development, saying 'Linux is not anti-AI' and 'If you don't like it, just fork it' - GIGAZINE
Related Posts:
