Law Firm AI Use Policy
.docx + .pdf · ~22 pages · adopt by partner voteA working AI use policy for any US law firm. Sections you actually need: data classification, tool tier matrix, prohibited uses, four-step verification protocol, court disclosure framework, engagement-letter language, supervisory protocol, training, incident response, audit log requirements, billing rules, and a 13-row sanctions ledger.
- 17 working sections + 4 appendices
- 2026 sanctions ledger (Mata, Lacey, ByoPlanet, Sullivan & Cromwell, ~13 cases)
- Vendor TOS decoded (Harvey, CoCounsel, ChatGPT current TOS)
- Engagement-letter language calibrated to ABA Op. 512 § 35
- State opinion crosswalk (CA, FL, NY, TX)
Legal AI Vendor Evaluation Checklist
.pdf · 60 questions · the engineer's versionA 60-question checklist for evaluating any legal-AI vendor before signing — and any vendor you've already signed. Calibrated for the technical clauses lawyers miss (data flow, model improvement traps, sub-processor inheritance) and the legal clauses engineers miss (Rule 1.6 posture, BAA, court acceptance, audit trail).
- 8 sections: Data Architecture, Training, Rule 1.6, Security, Contractual, Compliance, Court, TCO
- "Why it matters" explainer under every question
- Critical-question gates per section
- Decision matrix at the end
- Built from current Harvey + CoCounsel TOS analysis
// TERMS OF USE — PLAIN ENGLISH
These documents are free for any law firm or in-house team to use, customize, and distribute internally. They are not legal advice and do not establish an attorney-client relationship. Verify every authority against your jurisdiction before adopting. The documents reflect standards as of April 27, 2026; the area moves fast.
If these are useful to you and you want help implementing them — or you want a private AI workstation built specifically for your firm — that's what Knight CTO does. Schedule a 30-minute intake.
Found an error or want to suggest a section? Email boris@knightcto.com.