Updates to our AI Policy
September 5, 2026
Based on feedback and other things we observed since releasing our AI Policy, we have updated it with changes that are effective immediately.
Our policy on description/changelog text didn’t communicate what we intended. Previously, we stated that these can be generated by LLMs, but must be cleaned up by a human. This left it open for interpretation, as it’s unclear what “cleanup” is necessary. It could perhaps also be interpreted to encourage using an LLM to write a first draft - the opposite of our intention. This sentence was rewritten to: “Description/changelog text containing LLM speaking patterns will not be accepted.”
Added a paragraph regarding AI-generated UIs, since these often are poorly designed and contain LLM speaking patterns.
Added clarification that screenshots/videos should avoid showing AI-generated content, unless generating such content is a feature of the product. This includes content that may have incidentally appeared (for example, on a social media timeline) rather than being intentionally chosen.