Linux Kernel Drops Mandatory AI Disclosure Tag After Torvalds Critique
Linus Torvalds dismisses AI tagging requirements as ineffective and unnecessary for kernel development.
1Kernel Policy Shift
The Linux kernel community is officially deprecating the Assisted-by tag, which previously required contributors to disclose the use of AI coding tools. Proposed by Nvidia engineer Sasha Levin on October 9, the change follows a consensus reached at the Maintainers Summit to remove the documentation that mandated these specific commit message declarations.
2Arguments for Removal
Red Hat developer Jeff Layton advocated for dropping the requirement, citing poor compliance among contributors. He argued that the tag functioned primarily as free marketing for proprietary LLM companies. Furthermore, maintainers questioned the fundamental utility of collecting this data, noting that the policy failed to provide actionable insights for the kernel project.
3The Disclosure Debate
Laurent Pinchart of Ideas on Board opposed the removal, suggesting that the tag established necessary expectations for code transparency. He argued that even if some contributors bypassed the rule, it provided a mechanism for enforcement. Pinchart expressed concern that removing the tag would leave maintainers without a formal way to request disclosure.
4Torvalds' Sharp Rebuttal
Linus Torvalds dismissed the disclosure requirement as utterly stupid during a mailing list exchange on October 13. He sarcastically suggested that if tools must be disclosed, developers should also avoid compilers or text editors to prove their manual proficiency. He emphasized that understanding code matters far more than the tools used.
5Evolution of AI Sentiment
Torvalds revealed a shift in his personal perspective on AI during the Open Source Summit in Prague. While previously skeptical, he now utilizes AI for tasks outside his core expertise, such as interface design. He describes AI as a gateway for new programmers, noting its ability to improve the overall code base.
6Security and Code Quality
AI tools are actively contributing to the Linux kernel by identifying legitimate security vulnerabilities and flaws in legacy drivers that remained untouched for two decades. Despite these technical benefits, Torvalds noted that the influx of AI-generated contributions is creating significant stress for maintainers, who must now manage automated review systems.
7Automated Review Challenges
The Maintainers Summit highlighted that approximately three-quarters of recent discussions focused on mitigating the stress caused by AI tools. With the emergence of agentic review systems like Sashiko, some maintainers now expect patches to undergo automated checks. Torvalds acknowledged the irony of bots communicating with bots during the development process.
8Industry Divergence
While Linux moves toward integration, other projects are taking restrictive paths. System76 has banned AI-generated content from its COSMIC desktop, and NetBSD prohibited AI-generated code in 2024. NetBSD reported exceeding its donation target, suggesting that some segments of the open-source community prioritize human-only contributions despite the broader industry trend toward AI.