Linux app distribution platform Flathub recently added a new “Generative AI Policy” clause to its official documentation, clearly prohibiting two types of behavior: first, using AI tools to automatically generate, initiate, or automate the entire PR submission process, including manifests, metadata, patches, and build scripts; and second, apps that themselves contain AI-generated or AI-assisted code, documentation, or other content. Submissions that violate these rules can be rejected directly without any further review, and repeat offenders will be permanently banned. Additionally, the clause requires submitters to actively disable Copilot’s auto code review feature in their GitHub settings, to prevent AI tools from automatically initiating reviews in PRs. Flathub maintainer Bartłomiej Piotrowski added on Mastodon: „This was effectively an unwritten rule from the start — we’re just making it explicit now.“ The policy also notes that mature, well-maintained projects may be granted exemptions on a case-by-case basis.
This move makes Flathub one of the few major software distribution platforms known to explicitly list AI-generated content as a prohibited item in their submission rules, charting a different course from open-source projects like jqwik that actively resist AI agents — while those projects choose to fight back, Flathub intercepts at the source through policy. The Linux community has generally welcomed the change, noting that AI-assisted submissions have significantly increased reviewers’ workloads and brought in a large number of low-quality submissions with formatting errors or hidden bugs. Critics, however, argue that without reliable detection tools, the rule is hard to enforce effectively and mainly relies on submitters’ self-reporting.