FreeBSD survey backs AI disclosure and review rules for code

FreeBSD’s third status report of 2026 outlines active work across the project and ongoing debate over code produced with AI coding assistants. The survey suggests most respondents favour setting rules for AI use, while relatively few want AI-assisted code rejected outright.

The report contains 51 entries from five teams covering 16 projects, including Sylve, a graphical tool for managing FreeBSD virtualisation hosts, and AppJail, a system for distributing containerised applications. Of 140 survey respondents, 88% supported a project policy or guidance; 84% wanted AI use disclosed, and 78% called for mandatory human review. Sixteen per cent favoured rejecting AI-assisted code outright, while 75% expressed some or strong concern about copyright and licensing.