FreeBSD's stance on artificial intelligence-generated source code

In the recently published second quarter status report for 2025, the FreeBSD Project has provided an in-depth update on the project's progress. Here are some key highlights:

The report details significant updates to ports and packages, including a network driver for "Headphone-to-USB-serial" adapters and CPU pinning for Bhyve VMs. It also mentions changes in infrastructure, such as improvements to ZFS snapshots, which are now time-controlled.

One of the most notable developments is the ongoing investigation by the FreeBSD Core into the establishment of a policy for the use of Large Language Models (LLMs) in their open-source project. The report officially states that the Core is exploring this matter, but no decision has been made regarding the introduction of LLM-generated code.

This investigation comes amidst concerns about licensing issues and the origin of the content generated by LLMs. The outputs of these models are known to be derived from previously copied content from various sources without permission, which raises copyright concerns.

The resistance to the use of AI for code generation within the BSD community is palpable. Other open-source projects like NetBSD and Gentoo share similar sentiments, with NetBSD considering LLM-generated code "tainted" and Gentoo explicitly forbidding any code contributions created with LLM tools.

The QEMU project, too, has rejected code known or suspected to be generated by LLMs due to unclear license origin and possible non-compliance with the Developer Certificate of Origin (DCO).

The report also acknowledges the contributions made by the community, a testament to the project's collaborative spirit. Notably, the porting of FreeBSD to the PinePhone Pro is making progress.

Despite the controversies surrounding AI-generated code, the FreeBSD developers have decided not to integrate such code into the classic open-source operating system. This decision reflects the project's commitment to maintaining the highest standards of quality, copyright, and ethics.

The Linux Foundation, while generally allowing the use of AI tools, emphasises the responsibility of developers to ensure their work complies with the project's guidelines and licensing requirements.

In conclusion, the FreeBSD Q2 2025 status report offers an insightful look into the project's advancements, challenges, and the ongoing debate about the role of AI in open-source software development.