Module 6: Client Disclosure & Output Framing - DRAFT

When to Disclose AI Use

Disclosure is required when:

• AI substantially contributed to a deliverable.
• [Client Name] data was processed in AI tools.
• The client operates in a regulated industry (healthcare, financial, legal, etc.).

Recommended when:

• AI-supported research or analysis materially influences recommendations.

---

Framing That Works

Good examples:

• “I used Claude to accelerate the research and initial draft; I’ve reviewed and refined it, and the conclusions are mine.”
• “AI tools supported the analysis. All findings have been validated by our team.”

Avoid:

• “The AI said…”
• “The model thinks…”
• “This is just what ChatGPT/Claude gave me.”

---

The Four-Point Output Review

Before anything leaves RBA, Inc.:

1. Accuracy – facts, numbers, citations validated.
2. Confidentiality – no inadvertent leaks or cross-client references.
3. Bias / optics – no stereotyping, demographic assumptions, or exclusionary framing; see Module 7 for review guidance.
4. Fit for purpose – answers the right question for this client.

If any point fails, revise or escalate.

---

Internal Documentation

For each engagement, note:

• Whether AI was used.
• For what (research, drafting, analysis).
• Whether client data was processed and under what safeguards.