The latest FreeBSD report is here, and the AI debate is ongoing
Only a minority want to ban it. Plus plenty of updates on 16 projects
The FreeBSD project has released its third status report for 2026, showcasing ambitious work and a community divided on the integration of code generated by AI coding assistants. The comprehensive 51-page report covers 16 projects, with additional sections providing more detailed updates. Among the various developments, the team is working on a GUI for managing FreeBSD virtualization hosts, akin to Proxmox, and an app distribution system that can handle GUI applications, comparable to Snap, Flatpak, and AppImage on Linux.
However, the most significant discussion in the report revolves around the handling of code created by AI coding assistants. The FreeBSD Developer AI survey, conducted by Baptiste Daroussin, received 140 responses, revealing a mixed stance on AI usage within the project. While 88% of respondents favored adopting a policy on AI usage, either a formal guideline (49%) or light best practices (39%). Six percent opposed AI usage entirely, and another 3% deemed existing contribution rules sufficient.
Regarding AI-assisted code, 35% believed it should be disclosed, 25% advocated for stricter review, 22% proposed treating it like any other contribution, and 16% suggested rejecting it outright. Safeguards such as mandatory human review (78%) and disclosure (69%) were the most frequently requested measures, followed by provenance or licensing documentation (47%) and mandatory testing (39%).
Disclosure of AI usage was favored by 84% of respondents, with 48% wanting AI use to be disclosed at all times and 36% preferring disclosure when the involvement is significant. Additionally, 79% wanted AI-assisted commits to be marked in the commit log, and 56% preferred tags naming the tool. Concerns about quality, specifically correctness and subtle bugs (39%), licensing and provenance (26%), and maintainability (21%), were the main apprehensions, with 75% expressing concerns about copyright and licensing issues.
The consensus leaned toward accountability and transparency, with 91% and 78% respectively favoring these principles.
In terms of permissible AI uses, 65% of respondents would allow AI-generated documentation with human review, 54% would permit AI tools on project infrastructure with restrictions, and 29% would not permit it at all. This suggests that, similar to Debian, FreeBSD may take a cautious approach, allowing AI usage but encouraging responsible practices and disclosure.
Written by urgent.news from The Register Science's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
