nohaillo

Do not send AI-generated analysis to another person without adding your own human judgment.

The idea

An LLM can summarize, analyze, draft, and suggest. But its output is not your position. Sending it without context or opinion shifts the work of understanding, verifying, and deciding to the recipient.

Before sending AI output to a human

• Clearly label what was produced by an LLM.
• State your own view, recommendation, or decision.
• Explain what you want the recipient to review or do.
• Provide the original sources when they matter.
• Flag uncertainty, assumptions, and claims that need checking.

Agent output belongs with agents

If you only have an agent's output, send it to another agent for further processing. Do not make a human interpret an unreviewed, context-free AI attachment.

Humans should receive communication owned by a human.