Module 1: Accountability & the AI Code of Conduct - DRAFT
The Non-Negotiable Rule
Section titled “The Non-Negotiable Rule”You are accountable for every AI output you produce, share, or act upon.
Saying “that’s what [AI platform] gave me” does not transfer responsibility. It indicates a lack of professional ownership. Clients, colleagues, regulators, and courts will hold you—not the AI—responsible for errors, bias, hallucinations, fabricated citations, and privacy violations in your work product.
RBA, Inc. AI Code of Conduct
Section titled “RBA, Inc. AI Code of Conduct”-
Human primacy
AI never replaces human review and professional judgment. You own the deliverable. -
Truthfulness & non‑fabrication
You verify AI outputs before using them. If you cannot verify a claim, you do not use it. -
Transparency
You are candid about AI use and do not misrepresent AI work as purely human or vice versa. -
Data stewardship
Client and company data are handled under contract, law, and this policy—never pasted into unapproved tools. -
Least‑privilege use
Use the minimum AI capabilities, skills, and connectors needed. Extra permissions are extra risk. -
Incident responsibility
Suspected AI‑related issues (security, privacy, quality) are reported immediately via the incident process. -
Continuous learning
You keep current, complete annual training, and surface emerging risks observed in practice. -
Fairness & inclusion
You review AI outputs for bias, stereotyping, and accessibility concerns. AI-assisted work must not produce discriminatory outcomes.
Approved AI Platforms (Example Skeleton)
Section titled “Approved AI Platforms (Example Skeleton)”RBA, Inc. should maintain an up-to-date registry; this table shows the intended structure.
| Platform | Approved Use | Data Allowed | Notes |
|---|---|---|---|
| Claude | Deep analysis, writing, strategy | Internal & sanitized client | No raw PII/PHI unless contractually cleared. |
| Perplexity | Research, fact‑finding on public sources | Public/non‑confidential | Use for discovery, not for secrets. |
| [Others] | [Define] | [Define] | Maintain in AI tool registry. |
Using personal or unapproved AI accounts for RBA, Inc. work is prohibited.
Responsibilities by Role
Section titled “Responsibilities by Role”- Employees / Consultants: Apply this code in daily AI usage, from research to deliverables.
- Managers / Project Leads: Model good practice, enforce standards on engagements and projects.
- AI Governance & Risk Management Lead: Maintain tool registry, policies, guardrails, and training content.