Linus Torvalds, the creator and principal maintainer of the Linux kernel, has issued a firm response to developers who oppose the use of artificial intelligence in kernel development. In a recent post on the Linux Kernel Mailing List (LKML), Torvalds declared that anyone who cannot accept AI's role in the project is free to fork it—a direct reference to the open-source principle of creating a separate codebase. His statement arrives amid growing debate within the open-source community about whether AI-generated code should be allowed, especially after projects like the Zig language adopted strict policies against it.

Torvalds wrote, I realize that some people really dislike AI, but this is an area where I'm willing to absolutely put my foot down as the top-level maintainer. He emphasized that AI is a tool, much like compilers or version control systems, and that its utility is no longer in question. AI is a tool, just like other tools we use. And it's clearly a useful one. It may not have been that 'clearly' even just a year ago, but it's no longer in question today, he added.

Background on the AI Debate in Open Source

The Linux kernel has always been a bellwether for open-source practices. Since Torvalds first released the kernel in 1991, the project has grown to include thousands of contributors and millions of lines of code. Decisions made here often influence other projects. The current debate over AI stems from the rapid advancement of large language models (LLMs) and their ability to generate code, fix bugs, and automate backports. Critics argue that AI-generated code may introduce errors, lack attribution, or be misused for spam. However, supporters like Torvalds see it as an inevitable evolution of development tools.

In 2024 and 2025, many open-source projects began experimenting with AI assistants like GitHub Copilot, Claude, and GPT-4. By 2026, frontier models had matured significantly. Torvalds noted that these models are now capable of producing high-quality patches, especially for routine tasks like stable kernel backports. Greg Kroah-Hartman, maintainer of the Linux stable kernel, told ZDNET earlier this year that the quality of AI-generated security reports had improved dramatically. Months ago, we were getting what we called 'AI slop'—obviously wrong or low-quality reports. But then the world switched. Now we have real reports. All open-source projects have real reports that are made with AI, but they're good, and they're real, Kroah-Hartman said.

The SFC's Policy Statement and Reactions

The immediate trigger for Torvalds' post was a discussion about the Software Freedom Conservancy (SFC)'s recent policy statement titled When Using LLM-backed Generative AI Systems for FOSS Contributions. The SFC suggested that projects should support, not just tolerate, those who outright reject LLM-gen-AI systems. This provoked strong reactions among Linux kernel maintainers. Senior kernel developer Theodore Ts'o argued that such a policy could create impractical obligations. In his LKML response, Ts'o questioned: If someone rejects LLM-gen-AI systems, and the LTS kernel contains patches which are automated backported, and they object, are we bound to forswear the use of automated backport technologies? What if someone reports a bug with a kernel stack trace, and someone uses an LLM agent to analyze their bug report and find a fix? What does it mean to 'support someone who outright rejects the use of LLM-gen-AI systems' in that case? Ts'o concluded that it is not obvious the kernel community must bend over backward to accommodate such preferences.

Another top maintainer, James Bottomley, offered a pragmatic view: The contributor doesn't get to approve the tools the maintainer uses to assess and apply patches. If there's AI in there, and the contributor is an AI luddite, then the patch doesn't get applied (i.e., your right to ignore AI stops when it infringes others' right to use it). This aligns with Torvalds' own stance that technical merit should trump ideology.

Torvalds' Philosophy on Tools

Throughout his career, Torvalds has been known for his no-nonsense approach to development. He frequently reminds the community that Linux is built on technical excellence, not religious adherence. In his LKML post, he reiterated: In the kernel community, we do open source because it results in better technology, not because of religious reasons. And so we make decisions primarily based on technical merit. Not fear of new tools. This philosophy has guided the kernel through many transitions—from the adoption of Git (which Torvalds himself created) to the gradual inclusion of Rust as a second language alongside C.

On the topic of AI's limitations, Torvalds acknowledged that the technology can be somewhat painful but argued that the solution is not to ignore it. The solution is not to put your head in the sand and sing 'La La La, I can't hear you' at the top of your voice like some people seem to do, he wrote. Instead, he urged the community to work on making LLM tools help maintainers rather than causing them pain.

Practical Implications for Kernel Development

The integration of AI into Linux development is already happening. Automated backporting scripts use machine learning to identify which patches need to be applied to older stable kernels. Bug triage tools leverage LLMs to classify reports and suggest likely fixes. Even code review processes are being augmented with AI to check for common errors. These tools are not perfect, but they reduce the burden on human maintainers who are often overwhelmed by the volume of contributions.

For context, the Linux kernel receives around 10–15 patches per hour on average. The stable kernel team alone handles hundreds of patches per week. AI-assisted triage can save hours of manual work. Torvalds' stance ensures that these efficiencies are not lost due to ideological objections.

The contrast with projects like Zig is instructive. Zig's policy forbids any AI-generated code in its contribution process, citing concerns about quality and accountability. Torvalds implicitly criticized such blanket bans by arguing that they are technophobic. His position is that tools should be judged by their output, not by how they were produced.

What This Means for the Open-Source Community

Torvalds' declaration sends a strong signal to the broader open-source ecosystem. As the leader of the most influential open-source project, his endorsement of AI tools lowers barriers for other projects to follow suit. Smaller projects that were hesitant to adopt AI can now point to the Linux kernel as a precedent. Conversely, projects that choose to ban AI may find themselves isolated, especially if contributors prefer to work on projects where AI is allowed.

Several major distributions—including Fedora, Debian, and Ubuntu—already use AI to assist in package maintenance and bug fixing. The Linux Foundation, which oversees much of the kernel's development infrastructure, has invested in AI-driven security analysis tools. These developments suggest that AI is becoming deeply embedded in the open-source supply chain.

The debate is not entirely settled, however. Some developers worry about the environmental impact of training large models, or about the potential for AI to generate code that violates licenses. Torvalds did not address these concerns directly, but his emphasis on technical merit implies that such issues should be resolved through engineering—for example, by training more efficient models or improving license compliance checks.

Torvalds concluded his LKML post by reiterating that he will not force anyone to use AI, but he will very loudly ignore people who try to argue against other people from using it. This pragmatic stance reflects a long tradition in the kernel community: focus on what works, and let the results speak for themselves. As AI continues to evolve, it is likely that more and more aspects of kernel development will be assisted by intelligent tools. For now, Torvalds has made it clear that the door is open—and those who don't like it can always fork their own version of Linux.

Source:ZDNET News