Advanta is currently undergoing final system calibration ahead of launch. Selected infrastructure and experiences may still be in active refinement.

advanta

RC-7 · Risk Taxonomy 2026.1

Client confidentiality breach

Mechanism

Privileged or contractually confidential matter content reaches a third party via AI in a way that compromises privilege or breaches client obligations. Adjacent to Data leakage (RC-2) but distinct: this class names the downstream legal consequence (privilege waiver, protective-order breach, outside-counsel-guideline violation).

Evidence (what the Evidence Register holds)

Matter-level and client-level AI policy mapping; vendor approval with client-specific consent; matter intake protocol for AI restrictions; outside-counsel-guideline integration record.

Mitigation

Matter-level AI policy mapping; client-specific outside-counsel-guideline integration; matter intake screening for AI restrictions before tooling is deployed; vendor approval gated by client-specific consent.

Editorial Framing

Client confidentiality breach is the class whose consequence is highest — and whose mitigation is most procedural. Matter intake must screen for AI restrictions before tooling is deployed on the matter. Outside-counsel-guideline integration is non-negotiable: every client OCG with AI restrictions feeds the matter intake checklist.

Indicative Examples

  • Privilege waiver via AI processing of privileged content
  • Outside-counsel-guideline violation
  • Protective-order breach
  • Cross-matter information bleed via shared AI vector store