In a notable development reflecting profound shifts in the modern programming world, the Linux kernel project has officially announced the regulation and adoption of comprehensive rules governing the contributions of artificial intelligence tools in writing and developing code. The project published official guidelines allowing developers to use AI technologies and software across all aspects of kernel development, conditioned on a crucial requirement: placing full legal and technical responsibility on the human developer who submits the code. This step represents an explicit acknowledgment of the machine’s growing role in assisting humans while maintaining strict human oversight.
New rules for an advanced programming era
The new guidelines are now available within the project’s official documentation and were carefully drafted by Sasha Levin, a prominent kernel stability maintainer working at Nvidia, with strong and direct support from system founder Linus Torvalds. This regulatory policy emerged from a broad consensus reached during the 2025 Kernel Maintainers Summit held in Tokyo, Japan, before being officially integrated into the kernel repository last week.
Approval mechanism and legal liability
Under the new framework, AI agents are explicitly prohibited from adding the final sign-off tag to code. The rules emphasize that only the human element is authorized to certify the “developer’s certificate of origin,” the legal mechanism through which contributors guarantee that their code fully complies with the terms of the open-source system’s GPL 2.0 license.
The human developer is required to review all machine-generated code, ensure its compliance with licensing laws, and assume absolute responsibility for the contribution. Additionally, contributors must include a new attribution tag indicating that the work was assisted by external technologies, which must include the name of the AI tool, the model version number, and any specialized analysis tools used—for example: “Co-authored-by: Claude 3 Opus and Sachiko.”
Responsibility surpasses restriction
Both Torvalds and Levin made it clear in no uncertain terms that this policy is specifically designed to support well-intentioned developers while fully acknowledging that bad actors can simply hide their use of AI regardless of any imposed rules. As one report neatly summarized this approach: “You sign off on code generated by AI, so you own it. Legally, this has always been true, but now it is formally codified.”
These developments come at a time when programming tools based on large language models have become an integral part of developers’ workflows. Recent polls show that about 84 percent of developers today rely on these tools in their daily work. Kernel maintainers responded to this fast-paced reality by integrating an AI-driven code review system called “Sachiko” into their workflows. In this context, Greg Kroah-Hartman, a senior kernel maintainer, noted last March that AI-driven activities regarding system security and code review have seen a massive surge in recent weeks.
Broader implications for open-source projects
The approach taken by the Linux kernel is expected to serve as an ideal model that other open-source projects can emulate when facing similar questions and challenges. What is striking about this policy is its ability to bypass the fierce and ongoing debate over whether AI-generated code, pre-trained on copyrighted material, can be licensed under open-source licenses. Rather than diving into this legal quagmire, the project focused on the most important practical question: Who takes responsibility when things go wrong?
Although community discussions continue to highlight unresolved copyright concerns, the kernel team chose to enshrine the principles of transparency and human accountability as the guiding compass in this new programming era, emphasizing that AI is an enabling tool rather than a substitute for conscious human oversight.
The future of software development under artificial intelligence
Tech industry experts believe this regulatory step will open new horizons for accelerating the development of complex systems without compromising quality and reliability standards. Requiring careful human auditing ensures that security vulnerabilities or logical errors resulting from AI model hallucinations do not slip through. Thus, the Linux environment once again proves its high flexibility and ability to absorb the latest innovative technologies in a responsible and sustainable manner. Ultimately, human minds remain the ultimate guarantor of the security and stability of the technical systems the world depends on, no matter how advanced assistive machines become.
Frequently Asked Questions
Can artificial intelligence add the final sign-off signature to Linux code?
No, the new rules explicitly prohibit AI tools from adding the final sign-off tag. The human developer must review the code, assume responsibility, and approve it personally to ensure its compliance with open-source license terms.
What is the new assistance tag required when using AI?
Developers must use an attribution tag showing that the code was assisted by AI, and this tag must include the name of the AI tool, the model version number, and any specialized analysis tools used to complete the code.
How does the Linux project handle AI copyright issues?
The project bypasses the complex debate over whether AI-generated code violates copyright by focusing directly on transparency and human accountability, where the developer submitting the code assumes full legal responsibility for compliance with official licenses.
