FreeBSD discloses an AI working group, with 88 percent of surveyed developers backing formal usage rules
FreeBSD's Q3 progress report shows a survey of 140 developers in which 88 percent backed formal policies or guidelines and 91 percent held that humans remain responsible even when AI is used, with an AI working group formed to deliver a final report in Q4.
The FreeBSD project recently published its 2026 third-quarter progress report, covering 51 items across AI-assisted development, security compliance, virtualization management and desktop experience, with defining boundaries for AI participation in open source development among the quarter's key issues.
A survey of 140 developers found 88 percent support for formal policies or guidelines and 91 percent agree that humans remain responsible even when AI is used. For AI-assisted code, 35 percent want mandatory disclosure, 25 percent want stricter review than human-written code, 16 percent prefer outright rejection, and 65 percent support using AI-generated documentation after human review. The project has formed an AI working group and expects to deliver a final report in the fourth quarter, with no formal policy published yet.
On security, FreeBSD continues preparing for European Union Cyber Resilience Act compliance, refining vulnerability handling processes, and building software bill of materials tooling to track system components, dependencies and license information, with the underlying system design still under review. On experience, the project developed a graphical interface for the Bhyve hypervisor written in Free Pascal and Lazarus called Bhyvemgr, while Sylve, the unified system management platform for FreeBSD, continues to improve its web management interface with new storage options and data center clustering features added last quarter. The report indicates that the central challenge for open source projects bringing AI into their development workflow is not technical capability but accountability and code trustworthiness.