The Linux kernel has dropped a rule, implemented in January this year, that required contributors submitting code to the project to explicitly declare any usage of AI tools in that code’s creation.
At the start of this year, the Linux project added lines to its documentation noting that “When AI tools contribute to kernel development, proper attribution helps track the evolving role of AI in the development process… If you used any sort of advanced coding tool in the creation of your patch, you need to acknowledge that use by adding an Assisted-by tag.” Now, as reported by The Register, the project’s maintainers have decided the tag is no longer necessary, and it’s been removed from the relevant docs.
Why? Well, to hear Linux creator and lead maintainer Linus Torvalds put it in response to concerns raised about the tag’s removal by contributor Laurent Pinchart, because it doesn’t really communicate anything of value about the code it’s appended to.
Latest Videos FromPC Gamer
“I don’t understand this whole idiotic concept of ‘don’t use AI for X,'” Torvalds wrote in a post to the kernel mailing list. “The kind of expectations *I* have is that people do their best and I expect people to understand what they are doing. But that understanding has very little to do with what tools they use.”
In other words, there’s good code—which a creator knows well and understands—and bad code—which a creator has more or less barfed out—and both can be made with the help of AI tools. So in Torvalds’ view, there’s not really much value to forcing the disclosure of particular tools; it doesn’t help Linux kernel maintainers sift the bad code from the good. At least, that’s how I interpret Torvalds’ statement.
You may like
Torvalds compared rules restricting or banning AI use in coding (which, to be clear, is not what the rule under discussion did) to hypothetical rules that restrict other, more venerable tools. “Here’s a new rule just for you: don’t use compilers for any of your kernel submissions,” he suggested. “Because if you can’t show that you really understand what the machine is actually doing by writing your code in assembler, then I don’t want to see your code contribution. You clearly aren’t competent if you can’t write it that way.
“Do you not see how UTTERLY STUPID those kinds of rules would be? So no. We’re not going to make stupid rules that say ‘you can’t use tools for X'”.
It’s not a surprising statement from Torvalds, who has generally been pretty positive about the uses of AI when it comes to coding (while lamenting the number of slop reports the kernel team now has to wade through). I doubt arguments like this will end any time soon, though, even with the input of Linux’s benevolent dictator for life. Computing, and open-source especially, is still figuring out how to live in the world AI has brought us.