Add Generative AI disclosure policy to contribution guidelines and PR template by shlokgilda · Pull Request #3640 · chaoss/augur

shlokgilda · 2026-01-31T19:01:56Z

Description: Adds an AI disclosure policy to Augur, covering two files:

PR template: new "Generative AI Disclosure" section with a checkbox and three sub-questions (tools used, how, what was changed after review)
CONTRIBUTING.md: new "Generative AI Policy" section between "Signing-off on Commits" and "Keeping in sync" — covers requirements, rationale, scope, quality standards, copyright, and enforcement.
This PR closes Proposal: Add generative AI disclosure checkbox next to the DCO checkbox in the PR template #3371.

Notes for Reviewers: This is a draft for all maintainers to workshop. A few things worth discussing:

Wording of the three PR template sub-questions — is "What did you change after review?" the right framing?
Whether the "Engage as yourself" requirement is too prescriptive or about right
The enforcement section intentionally defers to the existing CHAOSS Code of Conduct rather than defining new consequences
I have added the references that I used to draft the policy in CONTRIBUTING.md. Is that the norm or no?
Commit trailers (Assisted-by:) and AGENTS.md are intentionally left out for now

References used: ada-url, Zulip, LLVM, Linux Foundation.

Signed commits

Yes, I signed my commits.

AI Disclosure: Used Claude to improve grammar.

… template Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>

MoralCode

Initial first pass review - i think we can maybe also refactor the may short middle sections into a different structure ( short paragraphs?) that says the same thing but may take up less space?

Unsure. will discuss over the coming days

Overall this is a great start!

MoralCode · 2026-02-03T14:36:46Z

.github/PULL_REQUEST_TEMPLATE/PULL_REQUEST_TEMPLATE.md


 ## Related issues/PRs
- This is a bulleted list of any related issues/PRs; leave blank if there are none
+- This is a bulleted list of any related issues/PRs; leave blank if there are none


we should maybe have this be a comment so that we dont have this text appearing in every PR if it is not edited

Good call, will wrap the placeholder text in comments so it doesn't clutter PRs.

I could wrap all placeholder texts in this file: "Changeset," "Notes," and "Related issues/PRs". Slightly out of scope of this PR but if it's okay, I can do it. LMK.

.github/PULL_REQUEST_TEMPLATE/PULL_REQUEST_TEMPLATE.md

MoralCode · 2026-02-03T14:44:18Z

CONTRIBUTING.md


+## Generative AI Policy
+
+You can use AI tools to help with your contributions. **You are the author, and you are accountable for the result.**


i worry that people will read this first sentence and then be like "ok cool, its allowed" and then turn their brains off and not read the rest of the policy

Suggested change

You can use AI tools to help with your contributions. **You are the author, and you are accountable for the result.**

We allow contributions that utilize generative AI tools provided they follow our policy of responsible AI use. This policy boils down to: **You, the author, are responsible for ensuring any contributions under your name**. These policies are primarily designed to ensure that the Augur project remains sustainable and healthy into the future.

I get the concern about people reading the first line and tuning out. Your version is a bit wordy though. How about something like: "You can use AI tools to help with your contributions, subject to the requirements below. You are the author, and you are accountable for the result." Keeps it concise but signals there's more to read?

CONTRIBUTING.md

Co-authored-by: Adrian Edwards <17362949+MoralCode@users.noreply.github.com> Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>

shlokgilda · 2026-02-03T16:11:14Z

Will look at consolidating the shorter sections once the other changes land. Might be easier to evaluate the length/structure after the content is finalized.

Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>

Consolidates the AI disclosure section into the actual default PR template (pull_request_template.md) and removes the unused PULL_REQUEST_TEMPLATE/ directory that wasn't auto-populating. Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>

MoralCode

added some feedback since i saw there were new commits here

MoralCode · 2026-02-16T20:10:51Z

CONTRIBUTING.md

+
+1. **Disclose AI use.** Complete the AI disclosure section in the PR template. This is mandatory. If no AI was used, check the "No generative AI" box — that's a valid answer, and it helps us distinguish from incomplete templates.
+
+2. **Review before submitting.** Read and test all AI-generated code or text before asking maintainers to review it. If you can't explain why a change works, it's not ready for a maintainer to review. If your changes are too long for you to read and review, that is a good indication that they are also too long for maintainers to read.


maybe a better way to title this is "human in the loop"/human responsibility

MoralCode · 2026-02-16T20:11:27Z

CONTRIBUTING.md

+Reviewing contributions takes real time and energy. Submitting unreviewed AI output shifts that burden onto maintainers — this is sometimes called an "extractive contribution."
+
+Disclosure matters because:


the transition from unreviewed to undisclosed here seems strange

MoralCode · 2026-02-16T20:13:39Z

CONTRIBUTING.md

+
+- **Respect.** Discovering a PR was unreviewed AI output after spending time on it feels like a waste. Telling us upfront lets us calibrate.
+- **Research.** Augur measures open source community health. Understanding how AI tools are used in our own community is directly relevant to our mission.
+- **Honesty.** Citing your tools is the same principle as citing Stack Overflow. It builds trust.


Suggested change

- **Honesty.** Citing your tools is the same principle as citing Stack Overflow. It builds trust.

- **Honesty.** Citing your sources is the difference between plagiarism and reusing/adapting ideas. You should cite the tools you use, whether that is Stack Overflow or an AI tool. It also helps build trust.

MoralCode · 2026-02-16T20:14:00Z

CONTRIBUTING.md

+
+This applies to all contributions: code, issues, PR comments, documentation.
+
+**Disclose:** code generation, test generation, documentation drafting, AI-assisted code review, substantial refactoring suggestions.
+


this seems a bit redundant with what has already been said above

Suggested change

This applies to all contributions: code, issues, PR comments, documentation.

**Disclose:** code generation, test generation, documentation drafting, AI-assisted code review, substantial refactoring suggestions.

MoralCode · 2026-02-16T20:15:54Z

CONTRIBUTING.md

+
+**Disclose:** code generation, test generation, documentation drafting, AI-assisted code review, substantial refactoring suggestions.
+
+This policy covers generative AI tools specifically. Non-generative tools like spell check, grammar correction, syntax highlighting, and basic IDE autocomplete are not in scope.


Suggested change

This policy covers generative AI tools specifically. Non-generative tools like spell check, grammar correction, syntax highlighting, and basic IDE autocomplete are not in scope.

This policy covers generative AI tools, especially (but not only) when used for contributing code, issues, PR comments, and documentation. Non-generative tools like spell check, grammar correction, syntax highlighting, and basic/deterministic IDE autocomplete are not in scope.

this combines the deleted part above to be more concise without repeating stuff.

MoralCode · 2026-02-16T20:16:49Z

CONTRIBUTING.md

+
+### Quality standards
+
+AI-assisted contributions are held to the **same bar** as everything else. Good code is good code regardless of how it was written. Bad code gets rejected regardless of how it was written.


we should probbaly soften this language so we arent accusing anyone with a closed PR of having "bad code"

…ment guidelines Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>

Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>

shlokgilda · 2026-02-16T21:49:11Z

Pushed an update addressing the new feedback:

Renamed "What we require" → "Your Responsibility"
Reordered requirements so "Review your own work" comes first, then "Disclose"
Rewrote "Why we ask" around the honesty/citation framing
Removed "Quality standards" section (redundant) and consolidated "What counts" into one paragraph
Softened enforcement language. No more "bad code gets rejected"

Also added a general ## Contribution Guidelines section above the AI policy based on our earlier conversation about setting baseline expectations for all contributions. Figured this was a good place to land it since we're already touching CONTRIBUTING.md.

Add Generative AI disclosure policy to contribution guidelines and PR…

de1dd42

… template Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>

shlokgilda mentioned this pull request Jan 31, 2026

Proposal: Add generative AI disclosure checkbox next to the DCO checkbox in the PR template #3371

Open

1 task

sgoggins added the usability Usability related issues (bugs, features, etc) label Feb 2, 2026

MoralCode reviewed Feb 3, 2026

View reviewed changes

shlokgilda and others added 3 commits February 3, 2026 11:08

Update CONTRIBUTING.md

5415be4

Co-authored-by: Adrian Edwards <17362949+MoralCode@users.noreply.github.com> Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>

Update CONTRIBUTING.md

beeb859

Co-authored-by: Adrian Edwards <17362949+MoralCode@users.noreply.github.com> Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>

Update CONTRIBUTING.md

2809592

Co-authored-by: Adrian Edwards <17362949+MoralCode@users.noreply.github.com> Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>

Update AI disclosure policy in PR template and contribution guidelines

d92a260

Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>

MoralCode added the devops Development Operations label Feb 8, 2026

shlokgilda added 2 commits February 15, 2026 13:10

Merge branch 'main' into fix/issue-3371-ai-disclosure-policy

fc5bdf4

MoralCode reviewed Feb 16, 2026

View reviewed changes

shlokgilda added 2 commits February 16, 2026 16:46

Revise AI disclosure policy in CONTRIBUTING.md for clarity and engage…

62fd9a8

…ment guidelines Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>

Refine AI output review prompts in pull request template for clarity

ba55f80

Signed-off-by: Shlok Gilda <gildashlok@hotmail.com>


		## Generative AI Policy

		You can use AI tools to help with your contributions. You are the author, and you are accountable for the result.

	You can use AI tools to help with your contributions. You are the author, and you are accountable for the result.
	We allow contributions that utilize generative AI tools provided they follow our policy of responsible AI use. This policy boils down to: You, the author, are responsible for ensuring any contributions under your name. These policies are primarily designed to ensure that the Augur project remains sustainable and healthy into the future.


		1. Disclose AI use. Complete the AI disclosure section in the PR template. This is mandatory. If no AI was used, check the "No generative AI" box — that's a valid answer, and it helps us distinguish from incomplete templates.

		2. Review before submitting. Read and test all AI-generated code or text before asking maintainers to review it. If you can't explain why a change works, it's not ready for a maintainer to review. If your changes are too long for you to read and review, that is a good indication that they are also too long for maintainers to read.

		Reviewing contributions takes real time and energy. Submitting unreviewed AI output shifts that burden onto maintainers — this is sometimes called an "extractive contribution."

		Disclosure matters because:

	- Honesty. Citing your tools is the same principle as citing Stack Overflow. It builds trust.
	- Honesty. Citing your sources is the difference between plagiarism and reusing/adapting ideas. You should cite the tools you use, whether that is Stack Overflow or an AI tool. It also helps build trust.


		This applies to all contributions: code, issues, PR comments, documentation.

		Disclose: code generation, test generation, documentation drafting, AI-assisted code review, substantial refactoring suggestions.


		Disclose: code generation, test generation, documentation drafting, AI-assisted code review, substantial refactoring suggestions.

		This policy covers generative AI tools specifically. Non-generative tools like spell check, grammar correction, syntax highlighting, and basic IDE autocomplete are not in scope.

	This policy covers generative AI tools specifically. Non-generative tools like spell check, grammar correction, syntax highlighting, and basic IDE autocomplete are not in scope.
	This policy covers generative AI tools, especially (but not only) when used for contributing code, issues, PR comments, and documentation. Non-generative tools like spell check, grammar correction, syntax highlighting, and basic/deterministic IDE autocomplete are not in scope.


		### Quality standards

		AI-assisted contributions are held to the same bar as everything else. Good code is good code regardless of how it was written. Bad code gets rejected regardless of how it was written.

Conversation

shlokgilda commented Jan 31, 2026

Uh oh!

MoralCode left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

shlokgilda commented Feb 3, 2026

Uh oh!

MoralCode left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

shlokgilda commented Feb 16, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants