diff --git a/AI-POLICY.md b/AI-POLICY.md
new file mode 100644
index 000000000..d1d7e8bb8
--- /dev/null
+++ b/AI-POLICY.md
@@ -0,0 +1,116 @@
+# AI-assisted contributions policy
+
+This document defines how artificial intelligence (AI) and machine-assisted
+tools may be used when contributing to Weston. It applies to all code,
+documentation, and derivative works under the Weston project.
+
+AI tools can help with productivity, but they cannot replace understanding.
+Contributors remain fully responsible for what they submit. The goals of this
+policy are to preserve code quality, maintain license compliance, and ensure
+that human accountability remains central.
+
+---
+
+## 1. Scope
+
+This policy applies to all Weston contributions:
+- Source code and headers
+- Documentation and comments
+- Commit messages, changelogs, and design notes
+- Derived configuration or data files included in the repository
+
+The policy covers both human-written and AI-assisted material that becomes part
+of the Weston source tree or its official documentation.
+
+---
+
+## 2. Human responsibility
+
+- Each contributor is responsible for the correctness, clarity, and license
+  compliance of their work.
+- AI tools may assist, but they do not author. The human submitting the change
+  must understand every part of it.
+- Contributors must be able to explain the logic, intent, and expected behavior
+  of any code they contribute.
+
+Do not submit unverified AI output. Submissions that the author does not
+understand will not be accepted.
+
+---
+
+## 3. Disclosure of AI use
+
+If a contribution includes material generated or heavily assisted by AI, this
+must be disclosed clearly in the commit message or merge request description.
+
+Use one or both of the following trailers:
+
+- AI-Assisted: yes
+- AI-Tool: <tool name, version>
+
+Minor or incidental AI use, such as spell checking or text reflow, does not
+require disclosure.
+
+Disclosure is for transparency, not judgment.
+
+---
+
+## 4. Review and maintainability
+
+- Reviewers treat AI-assisted changes like any other submission: they must be
+  correct, readable, maintainable, and consistent with Weston’s architecture.
+- If the AI-generated content is unclear, unidiomatic, or hard to maintain,
+  it must be rewritten before acceptance.
+- Code and documentation merged into Weston must remain understandable by humans.
+- "AI said so" is not an acceptable justification for a design or implementation
+  decision.
+
+---
+
+## 5. Licensing and origin
+
+Weston is licensed under the MIT License.
+
+All submitted content, including AI-assisted material, must be compatible with
+the MIT license and must not include any restricted or incompatible material.
+
+AI tools may produce content resembling third-party sources. It is the
+contributor's duty to ensure that no copyrighted or restricted code is copied
+or derived. When in doubt, rewrite manually or omit the section.
+
+---
+
+## 6. Weston data and model training
+
+Weston's source code, documentation, and related materials are licensed under
+the MIT license. External parties using Weston data to train AI models must:
+
+- Respect the MIT license and attribution requirements.
+- Avoid implying Weston's endorsement of any trained model.
+- Provide clear provenance and licensing information for derived models.
+
+Aggressive scraping of Weston infrastructure is forbidden. Use official
+mirrors or published tarballs when possible.
+
+---
+
+## 7. Violations and corrections
+
+If a submission violates this policy (for example, undisclosed AI use or
+license incompatibility):
+
+1. The maintainer may request clarification or rewriting.
+2. The change may be rejected or reverted.
+3. Repeated or intentional violations may affect contributor trust.
+
+---
+
+### TL;DR
+
+- Humans decide; AI assists.
+- Understand what you submit.
+- Disclose significant AI help.
+- Respect Weston (MIT) licensing.
+- Keep Weston readable and maintainable.
+
+---
diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index dc58f1c0b..3e0f9e0a2 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -26,6 +26,14 @@ Weston formerly accepted patches via `git-send-email`, sent to
 [tracked using Patchwork](https://patchwork.freedesktop.org/project/wayland/).
 New email patches are no longer accepted.
 
+### AI-Policy
+
+Based on similar policies from
+[Fedora](https://communityblog.fedoraproject.org/council-policy-proposal-policy-on-ai-assisted-contributions/)
+and from [OpenInfra](https://openinfra.org/legal/ai-policy) Weston has a
+dedicated [AI-policy](AI-POLICY.md) related to that. Please read that before
+opening an MR.
+
 Finding something to work on
 ----------------------------