Law firms

AI security for law firms: protect client confidences

Lawyers and staff want AI for research, drafting and review. AILeakShield helps your firm see which tools are in use, keep privileged material out of them, and show clients you have a control in place.

Sample data shown. Select the image to enlarge it.

What the ABA says about confidentiality

ABA Formal Opinion 512 applies a lawyer's existing duties to generative AI. As the ABA Litigation Section explains, the opinion concludes that lawyers should obtain informed client consent before using information related to a representation in prompts to self-learning AI tools, even in a firm's closed system, and that boilerplate in an engagement letter isn't enough. Check the rules in your own jurisdiction too.

Where confidences slip into AI

It's rarely a decision. It's a busy associate with a deadline.

Drafting

A contract or brief is pasted in to tighten the language.

Summarizing

Discovery, depositions or a client email are uploaded for a quick summary.

Personal accounts

Someone uses a personal AI login on a firm computer.

AI note takers

A meeting assistant joins a client call through a connected calendar.

Unapproved tools

A new research or drafting tool appears before anyone has reviewed its terms.

Client names

A matter name in a prompt can reveal who the firm represents.

How AILeakShield helps

Controls that fit how a firm works, without banning useful tools.

Privilege markers

Detects document markers such as "privileged and confidential" and "attorney work product" before they're sent.

Client and matter terms

Add client names, matter names and codenames as your own sensitive terms.

Approved tools only

Review each tool, approve the ones that meet your standards, and restrict the rest.

Vendor terms at a glance

See whether a vendor trains on input, a key question in the ABA's guidance on terms of use.

Rules for each group

Set different rules for practice groups and staff.

A record clients can ask about

Show which tools are approved and what was stopped, without storing the text.

How firms roll it out

Start by finding out what's already happening.

  1. See what's connected

    Start with the free Microsoft 365 report, with nothing to install.

  2. Watch for two weeks

    Deploy to managed browsers in monitor mode. Nothing is blocked.

  3. Decide and enforce

    Approve the tools you trust, restrict the rest, and turn on your data rules.

  4. Keep the evidence

    Decisions, overrides and sign-offs are recorded for your next review.

Frequently asked questions

Does this replace client consent?

No. Consent and professional judgment remain with your lawyers. AILeakShield helps you control which tools are used and what goes into them.

Will it catch all privileged content?

No. It looks for document markers and your own terms, so unmarked privileged text can get through. It works alongside training and your AI policy.

Can we still use legal AI tools?

Yes. Approve the tools your firm has reviewed. AILeakShield doesn't require you to ban AI.

Is there a law firm framework pack?

Not as a named pack. Firms typically turn on the Legal & Confidential rule, block credentials and personal data, and add client and matter terms.

Does anyone at AILeakShield read our prompts?

Blocked content is never stored, and other prompt text is cleared at the end of each day (UTC). Our privacy policy sets out the narrow exceptions, such as a security investigation or a legal requirement.

See what AI can reach today

Start with a free report of the AI connected to your Microsoft 365. It doesn't read any email, files or chats.