Privacy and human approval

The work helper must operate inside data policy and stop before actions that require accountable human approval. This principle matters because an AI system produces likely output from the context and instructions it receives; it does not automatically know the organization’s current facts, private policy, unstated intent, or acceptable risk.

Use the following sequence for this page: classify sensitive data before choosing a tool; minimize and redact inputs where possible; separate recommendation, draft, approval, and execution; keep logs and a recovery path for consequential actions. The sequence is a guide, not a ritual. Skip a step only when its question truly has no effect on the outcome, and strengthen it when mistakes would be costly.

Return to the opening case: A draft workflow can ingest customer tickets, generate refunds, update records, and email customers using one broadly permissioned account. Begin by rewriting the request as a small contract. Name the intended reader or user, the authoritative material, the operation to perform, the required output, and the review owner. If current information is required, identify where it will come from. If exact calculation or action is required, assign that step to a deterministic tool or an approved system rather than relying on prose generation.

A useful instruction could follow this shape:

Goal: help [reader] accomplish [outcome]. Use only [named sources or supplied material] for factual claims. Perform [specific operation] and return [format]. Mark missing information as TBD or ask a focused question; do not guess. Before the result is used, [named person or role] will check [criteria].

• Assuming redaction is complete because names were removed. This hides an important assumption or removes a review point. Replace it with an explicit rule, a source check, or a human decision.
• Approving in bulk without sampling outputs. This hides an important assumption or removes a review point. Replace it with an explicit rule, a source check, or a human decision.
• Giving the helper credentials beyond the task. This hides an important assumption or removes a review point. Replace it with an explicit rule, a source check, or a human decision.

Turn Privacy and human approval into a template you can reuse this week: job statement, required evidence, failure mode to catch, and reviewer. Store it next to your other work helper notes. When the job changes, rewrite only the job statement and re-run the same failure-mode list — do not invent a new workflow from scratch. Role of this page: Safety and governance.