Content policy
Also known as: usage policy · safety policy
A content policy is the set of rules defining what a product’s AI is allowed to produce or do.
- What topics are off-limits.
- What formats/outputs are disallowed.
- How refusals and edge cases should be handled.
Why write it down: a policy turns "be safe" (vague) into checkable rules. It's the reference for red teaming, evals, and user expectations alike.