# Pressidium Performance AI Policy

Generative AI tools are a regular part of many developers’ workflows. They can help contributors explore the codebase, develop and evaluate solutions, write or refactor code, create tests, and improve documentation.

However, using AI does not change the expectations we have for contributors. **Contributors are responsible for everything they submit, regardless of how it was produced**.

AI-assisted contributions should not create additional work for maintainers simply because AI was involved. PRs that appear to have been submitted with little review, understanding, or engagement from the contributor **may be closed without further review**.

## Acceptable Use

**You may use AI tools** as part of your development workflow.

When submitting AI-assisted work, you are expected to:

* **Carefully review and understand** the changes you submit
* **Test the changes** where appropriate and verify that they solve the problem they claim to solve
* **Be able to explain, debug, and take ownership** of the contribution during review
* Ensure the contribution follows the project’s **coding standards, contribution guidelines, and existing conventions**
* Keep contributions **focused** and **reasonably sized** so they can be reviewed effectively
* Ensure that submitted material complies with the project’s **licensing requirements**
* **Be upfront about AI usage** and briefly describe how AI contributed to the work

Ultimately, you should be comfortable putting your name behind the contribution just as you would if you had written it entirely yourself.

## Unacceptable Use

Do **not**:

* Submit code, tests, documentation, or other material that you have not carefully **reviewed** or do not **understand**
* Submit contributions that you **cannot explain, contextualize, justify, or debug** during review
* Use AI-generated output in a way that **shifts the work** of understanding, validating, or cleaning up the contribution **onto maintainers or reviewers**
* Submit AI-generated content that copies or closely reproduces external work without **complying with its licensing and attribution requirements**
* Post raw AI-generated responses as comments on issues, pull requests, reviews, or discussions. AI may be used to help with wording or proofreading, but **communication with other contributors should reflect your own understanding and intent**.
* Post unnecessarily long or verbose AI-generated responses that make review and discussion harder. Summarize relevant information in your own words and keep the conversation focused. **Maintainers should be discussing the contribution with you**, not effectively carrying on a conversation with an AI tool through you.

## Transparency & Attribution

If you use AI to prepare a contribution, you are expected to **disclose its use and briefly describe how it was used**. This applies to code, tests, documentation, as well as PR descriptions, issue reports, responses to code review, and other communication with maintainers and contributors.

You do not need to disclose genuinely minor uses that do not meaningfully affect the content, such as spellchecking, grammar corrections, simple autocomplete, or formatting changes.

**When in doubt, disclose it**. Transparency helps maintainers understand both how a contribution was produced and who they are communicating with during the review process. **Disclosure does not make a contribution less valuable or less likely to be accepted**, provided that you have reviewed and understood the work and can take responsibility for it.

Failure to disclose AI usage may result in a contribution being closed or rejected.