The open-source software community finally solved a massive argument about artificial intelligence. This week, the Linux kernel project established a clear rule allowing developers to use AI tools to write code. However, these developers must follow strict honesty rules. They cannot use the standard legal “Signed-off-by” label for computer-generated text. Instead, they must type a brand new “Assisted-by” tag on their submissions. This new policy legally forces the human developer to take 100% of the blame for any resulting bugs, crashes, or security flaws.
This practical decision ends months of pure chaos. A fierce argument peaked in January when Dave Hansen from Intel and Lorenzo Stoakes from Oracle fought over how the project should handle AI software. Linux creator Linus Torvalds eventually stepped in and shut the argument down completely. He called the push for outright bans pointless posturing. Torvalds look at artificial intelligence as just another everyday programming tool. He knows that bad actors will ignore the rules anyway, so the community should simply hold human developers fully accountable for the exact code they submit.
Until today, other major software groups have tried entirely different strategies. Over the past 2 years, groups that run Linux systems, such as Gentoo, and Unix systems, such as NetBSD, have enacted total bans on AI submissions. NetBSD leaders called the AI code legally tainted because nobody knows exactly who owns the original training data. Red Hat explained this legal nightmare late last year. They pointed out that developers legally certify they own their work through a Developer Certificate of Origin. Since chatbots are trained on millions of lines of open-source code under strict licenses, a developer using ChatGPT cannot guarantee they have the legal right to share that specific code.
Beyond the legal headaches, project leaders constantly fight a massive wave of pure garbage. The programming community simply calls this mess AI slop. People are trying to make a quick $500 by flooding containers with broken, computer-generated code. The creator of the cURL project actually had to shut down his reward program because he received too much fake code. Other groups, like tldraw, started automatically deleting outside contributions. Meanwhile, massive software projects like Node.js and OCaml spent hours arguing after users submitted broken 10,000-line AI patches that wasted everyone’s time.
Keeping AI usage a secret causes intense anger among volunteers. Late last year, an NVIDIA engineer and kernel maintainer named Sasha Levin faced massive backlash. He submitted a software patch for the 6.15 kernel update that a chatbot wrote completely from scratch. He never told anyone about the AI assistance. The code worked, but it cslowedthe system tdown Torvalds later admitted that the community failed to review the patch properly because it lacked an AI warning label. Programmers hate it when someone slaps their own name on a complex project they did not actually write.
The classic Doom video game community suffered a massive split over this same dishonesty. Last year, developers caught Christoph Oelckers using hidden AI code in the highly popular GZDoom project. When players and coders asked him for basic transparency, he brushed them off, telling them they could leave and start their own project. The community called his bluff immediately. The vast majority of the volunteer coders abandoned him and built a brand-new version called UZDoom, splitting the fan base into 2 groups.
These angry community fights show exactly why the new Linux kernel policy makes perfect sense. Programmers do not necessarily hate AI; they just hate the dishonesty surrounding free $0 coding tools. By forcing developers to use the new “Assisted-by” tag, the Linux kernel removes the emotion from the argument. Torvalds accept reality. Coders will use AI to work faster, and trying to ban the software is completely impossible. If the code works well, the project will accept it. But if a developer submits hallucinatory garbage that breaks the system, that human will face Linus Torvalds directly. In the programming world, no better deterrent exists.
