Torvalds says Linux is getting buried under repeat AI bug reports
He said the problem is not AI itself, but the pile of near-identical reports that stop short of sending a fix.
Linus Torvalds said Linux’s security queue has become far harder to manage after a wave of AI-assisted bug reports kept landing with the same problems attached. In a May 17, 2026 post on the Linux kernel mailing list archive, he wrote that “the continued flood of AI reports has basically made the security list almost entirely unmanageable” because different people are surfacing the same issues with the same tools.
The post came as Torvalds discussed the latest kernel release candidate. He said new drivers account for roughly half of the update, with GPU work making up a large part of that. The rest of the changes cover networking, core kernel work, filesystems, and architecture updates.
His criticism was not aimed at AI tools themselves. Torvalds has been open to using LLMs in software work, but he said the problem starts when a bug report stops at “I used AI and found this.” In his view, that kind of message does not add much when someone else has likely found the same issue already.
Torvalds said anyone who wants to help should do more than point out the bug. He urged people to read the documentation, write a patch, and add real value on top of what the AI found. He also suggested Linux may eventually need an automated filter for these submissions, and said it would be fine if that filter ended up being AI-based too.
Share your thoughts in the comments, and follow us on X, Bluesky, YouTube, Instagram.
