Thanks for sharing these. I think it’s important to be up front with an organization’s approach to AI and I am glad that you are sharing these for others to view. I do have some feedback/comments.
Mainly, they read as though human and AI are two wholly independent subsystems rather than an integrated team where the intent is coordination. For example, items 7 and 8 evoke the thought of using AI as a subroutine (invoke when necessary and dismiss).
Additionally, they read as though the human user is only in position to react to AI choices (once something has gone wrong) rather than proactively working with the AI to come with up with a solution — e.g., where they see the AI reaching a boundary condition or edge case.
There is also no mention of user control over automation or guidance of it, which is often an important part in system success (you do mention the system should learn, but that sounds largely after the fact and not active control).
For reference, my primary starting source when thinking about incorporating AI (or any automation) into a product is this book chapter by Christofferson and Woods: