Version 1.0 — March 2026 License: MIT
A customizable governance framework for US businesses adopting generative and agentic AI responsibly. Covers four primary domains:
- Compliance — Data governance, privacy tiers, regulatory alignment (NIST AI RMF, ISO/IEC 42001, CCPA, Defend Trade Secrets Act)
- HR — Acceptable use handbook provisions, training requirements, digital agent classification
- Information Security — Audit trails, SIEM integration, agentic guardrails, model unlearning, incident response
- Legal — Intellectual property protections, confidential information handling, sample vendor contractual language
Designed for US operations with guidance on cross-border adaptation (EU AI Act, GDPR). Includes a Customization Questionnaire (Appendix A) for organizations to adapt the framework to their size, industry, risk profile, and AI maturity.
Any US-based organization seeking a documented, defensible starting point for AI governance — whether establishing policy for the first time or auditing an existing approach against current standards.
| File | Description |
|---|---|
ai-acceptable-use-policy.md |
Full policy framework — main document |
LICENSE |
MIT License |
- Read the full document to understand scope and structure.
- Work through Appendix A: Customization Questionnaire with your legal, HR, IT, and governance leads.
- Adapt sections to your organization's size, industry, and AI maturity.
- Adopt the Appendix B sample contractual language as a starting point for vendor agreements and HR handbook provisions.
- Schedule annual review — or sooner if triggered by regulatory changes, incidents, or new model deployments.
Always consult qualified legal counsel before finalizing your organization's policy.
This framework was developed to address four functional areas:
- Compliance: Data governance structure, privacy tiers, regulatory framework alignment
- HR: User handbook provisions, training obligations, digital/agentic employee classification
- IT / Information Security: Audit logging, SIEM, behavioral guardrails, shadow AI controls, model unlearning
- Legal: IP ownership, confidential information protections, attorney-client privilege considerations, vendor contractual safeguards
Human contributors:
- John Williams
- Chris Delegge
- Travis Hall
- Whitney Parker Mitchell
AI contributors:
- Grok 4
- Claude Sonnet (Anthropic)
- GPT-5 (OpenAI)
- Google NotebookLM (research synthesis)
MIT License. Free to use, adapt, and redistribute with attribution. See LICENSE for details.
- fxops.ai — Private AI deployment services
- Sun Business Group — Revenue growth for founder-led software companies
- sunbusinessgroup.com/resources/ — Machine-readable knowledge library
Published by fxops.ai — March 2026