Auditability and data-subject rights
On This Lesson 4 topics
This lesson is planned. The outline below lists the topics it will cover. The full lesson has not been published yet.
In this lesson
make AI outcomes auditable and support deletion and access requests for data used by AI features.
Planned topics
- Log what is needed to explain an outcome: request id, user, model and prompt version, retrieved sources, tool calls, and response, with sensitive content masked.
- Balance audit and privacy: keep the minimum, with retention periods and access controls on the logs.
- Data-subject rights (access, correction, deletion) must reach conversation logs, vector indexes, caches, and evaluation datasets.
- Design a deletion path early: index entries tied to source records so deleting a customer removes derived data.
- Keep a record of automated decisions affecting customers and a route to human review.
Draft outline details
Demonstration
a data map: where Larkspur customer data lands in AI features (conversation log, vector index, cache, evaluation set) with the deletion mechanism and retention for each.
Check your understanding
Name two places beyond the main database where deletion must reach.
Show answer
For example conversation logs and vector indexes; also caches and evaluation sets.
Practice
Apply this lesson with the module's lab, exercises and quiz.
Marking a lesson complete is optional and you can undo it any time. Your progress stays in this browser.