QGIS, a widely-used open-source geographic information system, has enacted a strict moratorium on AI-generated code contributions through a new Enhancement Proposal (QEP428), according to discussion on the project’s GitHub repository.

The policy represents a significant tightening from QGIS’s earlier “human in the loop” approach to AI contributions. Project maintainers cited concerns that AI-assisted code submissions have strained the already overburdened reviewer community and undermined developer skill-building within the project.
One contributor described creating AI-assisted pull requests to investigate and fix bugs, but acknowledged the practice had broken “a trust contract between the reviewer and the contributor.” They noted that AI tools create an illusion of competency while preventing developers from building genuine expertise: “by making this kind of ease available, contributors using AI learn very little/nothing and don’t build up their skills on the project.”
Project leadership acknowledged the new policy’s restrictiveness is intentional. According to discussions, the approach involves “starting with a closed door, which we can then selectively open a little if/when it’s safe and proper discussion has been made.” Maintainers indicated that specific exceptions—such as AI use for IDE autocomplete or language translation—may be reconsidered on a case-by-case basis following further community deliberation.
The previous policy had not functioned as intended in practice. Of approximately 50 pull requests tagged as AI-assisted, roughly 40 were closed without merging. The policy change also reportedly contributed to at least one highly valued contributor leaving the project.
Some community members raised concerns about the policy’s ambiguity, questioning which tools and workflows would be prohibited—from GitHub Copilot to machine-learning-based IDE autocomplete features to using AI for brainstorming. The proposal has since moved into a voting stage, with maintainers emphasizing that the moratorium is not intended to be permanent.
Key facts
- QGIS implemented a complete moratorium on AI-assisted code contributions through QEP428
- The policy represents a reversal of the project’s previous “human in the loop” AI contribution approach
- Concerns cited include strain on reviewers and reduced skill development among contributors
- Approximately 40 of 50 AI-assisted pull requests were closed without merging under the previous policy
- Project leadership framed the ban as intentionally restrictive, with potential selective reopening after further discussion
- The proposal has moved to a voting stage after discussion period concluded
