LLVM project adopts "human in the loop" policy following AI-driven nuisance contributions • DEVCLASS

The LLVM compiler project has banned AI-generated code contributions without human approval due to increasing low-quality submissions. The new "human in the loop" policy requires contributors to fully review and understand AI-assisted code, not just glance over it. Contributors must be able to answer questions about their submissions without referencing the AI tool. The policy also prohibits AI use on "good first issue" tasks meant for learning. This follows similar moves by other major open source projects like cURL, which closed its bug bounty program due to AI-generated spam. Some community members argue the policy is too permissive and should ban AI contributions entirely.