The privilege problem with public AI

Pasting client material into a consumer AI tool can constitute disclosure to a third party — and courts weigh whether a firm took reasonable steps to protect confidentiality. Shadow AI use by associates is now a privilege risk, not just an IT one.

A governance framework

  1. An acceptable-use policy that names approved tools and prohibits client data in unapproved ones.
  2. Enterprise AI with contractual data protections (no training on your inputs).
  3. Data classification so sensitive matter material is handled correctly.
  4. Training, so the rules are understood, not just published.

Enable, don't ban

Prohibition drives AI underground. The firms that win deploy governed tools that accelerate research and review while keeping the privilege perimeter intact. See how we work with legal practices.