Why open source projects are banning AI contributions
The growing divide between AI hype and software engineering reality
A growing number of open source projects are banning AI-assisted contributions, citing quality concerns and the burden on maintainers. A review of 120 projects found 37 with total AI bans, including GCC, QEMU, and Zig, while Debian votes on a similar policy. The author argues that LLMs are deceptive, produce significant errors, and that information asymmetry leads juniors to submit flawed code, wasting senior developers' time. Despite benchmarks improving, real-world reliability remains low, with experts like Greg Kroah-Hartman noting that at least a third of AI-generated results are wrong or harmful.
Even with the best of the current and next generation tools, at least 1/3 of the results they generate are flat out wrong or harmful.