Flathub has banned AI-assisted submissions across its Flatpak repository, but critics argue the policy falls hardest on open-source developers because proprietary applications cannot be inspected for AI use.
Flathub has introduced a sweeping ban on AI-generated and AI-assisted content across its Flatpak application repository, prohibiting the use of tools such as ChatGPT, GitHub Copilot, and other AI assistants throughout the submission process. Effective from 29 May 2026, the policy covers application code, BaseApps, extensions, build scripts, manifests, metadata, documentation, and even pull request text. Violating submissions can be rejected without review, while repeat offenders risk a permanent ban.
The move has ignited debate across the open-source community, with critics arguing that enforcement disproportionately affects open-source developers. Because open-source code is publicly available, reviewers can inspect submissions for signs of AI generation, including inconsistent coding styles, hallucinated function names, prompt fragments, or mismatched documentation. Proprietary applications hosted on Flathub, however, cannot be technically inspected, creating what critics describe as an enforcement imbalance.
Flathub maintainer Bart Piotrowski said the decision followed a surge in problematic AI-assisted submissions and increasingly hostile interactions with rejected contributors. “I’m tired,” he wrote, describing some submitters as acting as though they were “bestowing their brilliant software upon us idiots who are rejecting it.”
The policy arrives amid broader open-source concerns over AI-generated contributions. Linux creator Linus Torvalds recently said the kernel’s security reporting channel had become “almost entirely unmanageable” due to duplicate AI-generated bug reports, while cURL creator Daniel Stenberg ended the project’s bug bounty programme after AI-generated submissions reached 20% of reports.
The ban also reflects growing concerns over software supply-chain risks, licence conflicts, copyright uncertainty, and AI-generated code quality, as the Linux ecosystem remains divided between outright prohibition and disclosure-based approaches.
