diff --git a/.github/pull_request_template.md b/.github/pull_request_template.md
index 03c0b8a6c..a62f8763f 100644
--- a/.github/pull_request_template.md
+++ b/.github/pull_request_template.md
@@ -3,3 +3,7 @@
 - [ ] Update docstrings/[documentation](https://github.com/openforcefield/openff-toolkit/tree/main/docs), if applicable
 - [ ] [Lint](https://docs.openforcefield.org/projects/toolkit/en/stable/users/developing.html#style-guide) codebase
 - [ ] Update [changelog](https://github.com/openforcefield/openff-toolkit/blob/main/docs/releasehistory.md)
+
+See our [AI Policy](https://github.com/openforcefield/openff-toolkit/blob/main/AI_POLICY.md). Please indicate whether LLMs or other AI-powered tools (beyond simple IDE use cases) were used in this contribution:
+- [ ] No
+- [ ] Yes: [If yes, please provide details here]
diff --git a/AI_POLICY.md b/AI_POLICY.md
new file mode 100644
index 000000000..b898d9843
--- /dev/null
+++ b/AI_POLICY.md
@@ -0,0 +1,28 @@
+# AI Policy
+
+## OpenFF Ecosystem Policies
+
+The Open Force Field Initiative reserves the right to reject contributions if we judge that they do not follow the spirit of these principles.
+The following policies hold for all repositories in the OpenFF ecosystem:
+
+1. Repository Maintainers always reserve the right to make exceptions to any of the following rules on a case-by-case basis.
+2. LLMs use must be always disclosed. At minimum in the PR description, but preferably reflected in a “Co-Authored-By” line in the commit.
+3. LLMs may always be used for translation (disclosing LLM use is important here, so that maintainers can be more considerate when encountering translation inaccuracies).
+4. All content must have a human identified as being accountable for the contribution. The human takes responsibility for the contribution and stands by its correctness. No agent-only commits.
+5. Code authors must retain full authorship for any contributed code and contributions must be covered by the [MIT license](https://github.com/openforcefield/openff-toolkit/blob/main/LICENSE), and [Developer Certificate of Origin](https://developercertificate.org/).
+
+## Policies specific to this repository
+
+The following policies are specific only to this repository, and decided by the repository's maintainer(s).
+
+In this repository, it is either acceptable (✅) or not acceptable (❌) to use LLMs for:
+
+- ✅️ Generating content that only the author sees, but affects their work (searching repositories, asking questions, finding bugs, self-review, linting).
+- ✅️ Generating code that is shared, but not committed to the code base (e.g. Minimally Reproducible Examples in issue text).
+- Generating content which the human author has:
+
+|                                                              | Code | Docs | PRs & Issues | Replies to humans |
+|--------------------------------------------------------------|:----:|:----:|:------------:|:-----------------:|
+| Read thoroughly and understands every line before submission |  ✅️  |  ✅️  |      ✅️      |         ❌         |
+| Checked for obvious issues before submission                 |  ❌   |  ❌   |      ❌       |         ❌         |
+| Not reviewed before submission                               |  ❌   |  ❌   |      ❌       |         ❌         |
diff --git a/docs/releasehistory.md b/docs/releasehistory.md
index 2bd396cc9..1ef4d104d 100644
--- a/docs/releasehistory.md
+++ b/docs/releasehistory.md
@@ -25,6 +25,9 @@ Releases follow the `major.minor.micro` scheme recommended by [PEP440](https://w
 
 ### Improved documentation and warnings
 
+- [PR #2248](https://github.com/openforcefield/openff-toolkit/pull/2248): Add initial AI policy.
+
+
 ## 0.19.0
 
 ### Behavior changes
diff --git a/openff/toolkit/topology/molecule.py b/openff/toolkit/topology/molecule.py
index d50301b83..862fbf30f 100644
--- a/openff/toolkit/topology/molecule.py
+++ b/openff/toolkit/topology/molecule.py
@@ -1170,7 +1170,14 @@ def to_dict(self) -> dict:
         # https://mypy.readthedocs.io/en/latest/typed_dict.html#typeddict
         molecule_dict: dict[
             str,
-            str | bytes | dict[str, Any] | list[str] | list[bytes] | list[HierarchyElement] | list[dict[str, Any]] | None,
+            str
+            | bytes
+            | dict[str, Any]
+            | list[str]
+            | list[bytes]
+            | list[HierarchyElement]
+            | list[dict[str, Any]]
+            | None,
         ] = dict()
         molecule_dict["name"] = self._name