+1 to Aman's original proposal and +1 to comments by Victor, Sujan, and Ádám. Especially that the debate is over about using AI or not as it's already in use: It's just a matter of degree of usage and the boundaries we set.
Aleks, I'm still digesting the messages you just sent in this thread but I guess I'd just add I'm generally against AI-generated documentation (especially when it is verbose and "comprehensive"), but some of the shorter sections of technical documentation I've generated (and proofread) turn out fine. To me it comes back to those top three rules above from fory's policy (below). Agreed that prompt➡output reproducibility is more or less nonexistant. I think the apache/fineract repo doesn't need an entire AI_POLICY.md file. I do like the first few bullets in fory's <https://github.com/apache/fory/blob/main/AI_POLICY.md>, so I propose <https://issues.apache.org/jira/browse/FINERACT-2800> we instead just add a short section to our existing contributor's guide <https://github.com/apache/fineract/blob/develop/CONTRIBUTING.md>. Something like this: ## AI Policy 1. AI tools MAY assist contribution work. 2. AI tools MUST NOT replace contributor accountability. 3. The human submitter is responsible for correctness, safety, performance, and maintainability of all submitted changes. For more ideas on properly submitting AI-assisted contributions, see https://github.com/apache/fory/blob/main/AI_POLICY.md While we're improving CONTRIBUTING.md, we should also 1. remove the incorrect "How to download Gradle wrapper" guide (we do bundle the gradle wrapper in the source tarball) 2. update one Java 21 reference to Java 25 (I missed that during FINERACT-2734 <https://issues.apache.org/jira/browse/FINERACT-2734>)
