Outcomes

- Write a simple AI usage policy that engineers will actually follow - Build a data classification checklist (what can/can’t go into models) - Set up least-privilege access + audit trails - Create an incident response plan for leaked keys or data