AI Use Policy
Rules and human-control requirements for AI-assisted work in Lawie.
Effective 2026-07-14 · Owner: Legal and Privacy
Assistance, not authority
Lawie AI can retrieve, classify, compare, summarize, draft, and propose workflow actions. It does not become the customer’s lawyer, decision maker, signatory, or legal authority. Consequential legal, financial, employment, compliance, or rights-affecting decisions require qualified human review.
Verification duties
- Open and verify cited sources; a citation shows provenance, not that every conclusion is correct.
- Treat confidence and evidence-strength indicators as decision support, not guarantees.
- Do not approve or communicate a material conclusion when required evidence is missing, stale, outside scope, or contradictory.
Prohibited AI use
- Illegal surveillance, doxxing, stalking, intimidation, discrimination, or circumvention of access controls.
- Fabricating evidence, citations, identities, signatures, or records; concealing AI use where disclosure is legally required.
- Fully automated high-impact decisions where law or professional duties require meaningful human review.
- Attempts to extract another tenant’s data, system prompts, credentials, or restricted model or security information.
Data handling and feedback
Users must submit only data they are authorized to process and should minimize sensitive content. Feedback should identify relevance or quality without adding secrets or privileged material unless the approved support channel and contract permit it.