Linux Drops the Mandatory AI Tag: Torvalds Bets on Competence, Not Disclosure

Direct answer

Linux kernel maintainers agreed to retire the mandatory "Assisted-by" tag that required disclosing AI tool use in submitted patches. The decision came at the Linux Kernel Maintainers Summit on October 8, 2026 in Prague: the next day, developer Sasha Levin sent the patch removing the requirement from the documentation, documentation maintainer Jonathan Corbet applied a revised version on October 14, and the change landed in the development tree for Linux 7.4 by October 21. Linux creator Linus Torvalds backs the change: the problem is not using AI, it is submitting code you do not understand.

Why the tag failed

The line that sums up the change

Torvalds compared banning AI to banning text editors, and asked whether it would not be absurd to demand that someone prove they can write a perfect commit message unaided. His line, as covered: he wants "competent people using AI as a tool" and expects people working on the kernel to "understand what they are doing". At Open Source Summit Europe, he also called vibe-coding a joyful way into programming and a gateway to the craft.

The lesson for your business

  1. Judge the deliverable, not the tool: a policy demanding disclosure of every AI use sounds rigorous and ends up as dead paper if nobody complies. Review is what protects you.
  2. Accountability with a name on it: whoever delivers work must be able to explain and defend it. That rule holds for code, for a financial report and for a sales proposal.
  3. The tag does not replace review: if your process depends on someone checking a box to know whether something is trustworthy, your process is the checkbox. Change the process.
  4. There is no industry consensus: while Linux retires the tag, projects like NetBSD have banned AI code since 2024. There is no single answer: pick a policy and back it with real review.

The honest note

Related articles

Frequently asked questions

Did Linux ban AI-generated code?

No. The opposite: it retired the requirement to disclose its use. It neither mandates nor forbids AI tools; it judges them on quality.

What was the "Assisted-by" tag?

A line in the commit message, in force since January 2026, asking contributors to name the AI tools used to create the patch.

Why was it retired?

Because almost nobody used it (about 10% of commits) and because it did not say which part of the work AI did. A label without context does not help review.

What is still mandatory?

That the submitter understands and can explain the code, and that the origin sign-off comes from a human, never from AI.