The Linux kernel project has officially reversed its stance on artificial intelligence transparency by eliminating the mandatory requirement for developers to disclose when AI tools are used to write or assist in writing code. The decision, finalized following discussions at the Linux Maintainers Summit held on October 8, 2026, in Prague, Czech Republic, marks a significant shift in how one of the world’s largest open-source projects integrates modern development technologies.
Under the previous guidelines, contributors submitting AI-generated or AI-assisted code were required to include an “Assisted-by” tag within their commit messages, signaling that the code was not written entirely by a human hand. With the rule change, that documentation requirement has been scrapped, and developers are no longer obligated to provide extra declarations regarding their use of AI programming assistants.
The policy shift sparked immediate debate among open-source contributors, prompting Linux founder Linus Torvalds to address the issue directly on the project’s mailing lists. Responding to developers who argued that dropping the disclosure requirement might weaken project oversight over machine-generated code, Torvalds forcefully defended the integration of AI tools, calling opposition arguments “extremely stupid.”
Torvalds argued that the core issue is not what tools a developer uses, but whether they fully understand the code they are submitting. To illustrate his point, he suggested that demanding the avoidance of AI tools on conceptual grounds is akin to requiring developers to abandon compilers and write directly in assembly language to prove they comprehend underlying computer mechanisms. He further remarked that, by that same logic, text editors could be banned because backspace keys help conceal typing mistakes, ultimately rendering such arbitrary rules meaningless.
Rather than fixating on the origin of the code, Torvalds emphasized that his primary concern remains the contributor’s comprehension of what the code actually does and their ongoing responsibility for its maintenance. In his view, tools do not inherently dictate code quality; human understanding and accountability are the definitive metrics.
Over recent months, Torvalds has frequently voiced pragmatic and positive views regarding AI programming assistants. He has asserted that AI coding tools are here to stay, capable of boosting development efficiency and adding enjoyment to the software engineering process. Torvalds has also acknowledged using AI himself to help interpret complex software patches, and he recently utilized Anthropic’s Claude Code tool to generate implementation code, testing routines, and initial documentation for a kernel optimization task, while personally handling final code review and performance verification.
Supporters of the decision to scrap the disclosure tags pointed out that the prior rules suffered from limited enforcement, as some developers routinely failed to label AI usage, rendering the requirement largely symbolic. Additionally, some maintainers noted that the mandatory tag functioned effectively as free promotion for commercial AI vendors without providing tangible benefits for the kernel project.
Conversely, critics of the policy change voiced lingering concerns. These developers argued that even if compliance is imperfect, formal disclosure requirements communicate project behavioral expectations and offer code reviewers valuable contextual references. Without the tag, identifying whether a given contribution relied on automated assistance may become considerably more difficult.
By phasing out the “Assisted-by” label, the Linux kernel project has clarified its operational stance: the community will concentrate its oversight on code quality, functional correctness, and developer comprehension rather than policing the tools used during creation. This positions Linux among the most open-source projects taking an unrestrictive approach to AI-assisted software development.
