NetBSD Bans AI-Generated Code

NETBSD Foundation introduced updated rules amending the tree of the initial texts of the project. In the new rules, a item that prohibits the inclusion in the NETBSD code generated using AI tools based on large language models, such as ChatGPT, GITHUB COPILOT and CODE LLAMA, without preliminary written approval by the Core Team team. The reason is mentioned uncertainty in the field of copyright and inconsistency with the policy of licensing NetBSD.

Fear is associated with the fact that AI systems are trained on a large array of information, among other things, including the code protected by copyright and distributed under different licenses. When generating the code of AI systems, these features are not taken into account and potentially the result of AI can be considered as the creation of a derivative work from the code that was used in the training of the model and spreads under certain licenses.

When teaching a model on a code with a license requiring a notification of authorship, this requirement generated by the AI ​​tools is not formally fulfilled, which can be considered as a violation of most open licenses such as GPL, MIT and Apache. There may also be problems with licensed compatibility when inserting into projects under permissive licenses, such as BSD, code generated using models trained on code with copy-license.

/Reports, release notes, official announcements.