State bar guidance on AI and confidentiality.
A working summary of the ethics opinions and standing-committee guidance most relevant to mid-market law firms post-Heppner. We refresh this page on the first business day of each month.
- California
Cal. Standing Comm. on Prof. Resp. Op. 2026-201
Affirms that consumer-tier AI providers are non-lawyer third parties for purposes of Cal. Rule 1.6. Imposes an affirmative duty on supervising attorneys to verify technical safeguards. Heppner cited approvingly.
Privil alignmentShips a California policy template aligned to Op. 2026-201, including a supervisor-attestation export.
- New York
N.Y. State Bar Ass'n Ethics Op. 1265 (2026)
Holds that the use of consumer AI without contractual confidentiality protections violates N.Y. Rule 1.6. Specifies that enterprise-tier agreements meeting enumerated criteria are permissible.
Privil alignmentEnterprise-tier routing satisfies the criteria identified in Op. 1265. Routing decisions are logged and exportable.
- Florida
Fla. Bar Op. 26-3 (Apr. 2026)
Requires written informed consent from the client before any use of generative AI in matters involving the client's confidential information. Provides model consent language.
Privil alignmentAudit log includes per-matter consent metadata; supports linking the firm's intake forms to AI-use authorizations.
- District of Columbia
D.C. Bar Legal Ethics Op. 391 (2026)
Adopts the Heppner reasoning by reference and adds specific guidance on government-tier AI used in defense of regulated entities. Requires audit-trail retention for the duration of the representation plus seven years.
Privil alignmentAudit log retention policies are configurable per matter and meet the seven-year requirement out of the box.
- Illinois
ISBA Op. 26-04
Aligns with Heppner. Notes Illinois-specific data residency considerations for matters touching state-regulated industries (insurance, banking).
Privil alignmentSupports US-region-only routing and US-only audit log storage by default. Per-state residency policies configurable.
- Texas
Tex. Op. 700 (Public comment closed Apr. 2026)
Draft opinion would adopt the Heppner reasoning with one modification: a 'qualified provider' safe harbor for AI vendors meeting specified contractual and technical criteria.
Privil alignmentPrivil is on the working group for the Texas safe-harbor criteria. Updates will be reflected in the Texas policy template upon final issuance.
This page is informational and does not constitute legal advice. The text of each cited opinion controls. Material changes are noted in our monthly compliance digest.