Linux kernel maintainers drop requirement to disclose AI assistance
Linux kernel maintainers have agreed to remove the requirement for contributors to disclose AI assistance using an “Assisted-by” tag. Linus Torvalds said contributors should understand the code they submit, regardless of which tools they used, as AI-generated code has become more capable.
The disclosure proposal emerged in July 2025 and entered kernel documentation in January 2026. Supporters of dropping it say it is inconsistently followed, may advertise proprietary AI products and collects information of unclear value; opponents argue disclosure sets expectations and helps maintain accountability. Concerns remain about errors, security, copyright and whether developers can explain code produced with AI tools.
- Kernel contributors will no longer need an “Assisted-by” tag.
- Torvalds says understanding submitted code matters more than tool disclosure.
- Maintainers remain divided over transparency and accountability.
New here? Start with this
The Linux kernel is the core software that manages a computer’s hardware and lets other programs run. It is developed collaboratively by programmers around the world, who propose changes that are reviewed before they are added.
Linus Torvalds is the kernel’s lead maintainer, and other maintainers review contributions in their areas. The debate over AI assistance concerns how contributors should explain the tools used to produce code, alongside the expectation that they understand and take responsibility for what they submit.
The issue matters because the kernel is used in many devices and systems, so mistakes can affect reliability and security. Disclosure can help reviewers understand how code was produced, while requirements that are hard to apply consistently may offer limited information.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Removing the tag keeps review focused on whether contributors understand and can defend their code, regardless of how it was produced. A disclosure rule that is inconsistently followed may create unreliable records, promote particular tools and gather information without a clear use; code review and testing can address quality and security concerns directly.
The case against
A disclosure tag can help reviewers understand how code was produced and decide where to probe for errors, security issues or licensing concerns. Even if imperfectly followed, a clear expectation may encourage contributors to take responsibility for AI-assisted work and give maintainers useful context as these tools become more capable.
Read the full article at the source →
Originally published by The Register as “Linux kernel devs drop obligatory confession of AI tool usage”.