QuestionLegal ServicesAI GovernanceImplementation

How should a law firm assess AI use with privileged client material?

16 October 2026
Answered by Rohit Parmar-Mistry

Short answer

Before using AI with privileged or confidential matter material, a law firm should document the task, information flow, client and engagement restrictions, access, retention and human review. If key facts remain unclear, pause that use until the firm has evidence to assess it.

Detailed answer

There is no reliable blanket answer based on an AI product label alone. A law firm should assess the particular task, material and service, record what it knows, and identify who is accountable for the decision. This is an operational assessment, not legal advice or a conclusion about privilege.

Define the task and material

Start by recording what the AI would do and what information it would receive. Drafting from public material differs from summarising a privileged advice note or searching across matter files. Note the purpose, information categories, people who need access, and any relevant client instructions or engagement terms.

Map the information path

Establish what prompts, files, outputs and logs the service processes; which providers or people may access them; how long they are retained; and what deletion and model-improvement terms apply. Check current product settings and contract terms rather than relying on a general product description. Confirm how matter permissions are applied and whether information could cross client or matter boundaries.

Compare the available evidence with the firm’s confidentiality commitments, client instructions, engagement terms and internal policies. If a material fact is unknown, record it as an open question rather than treating uncertainty as proof that the use is safe or unsafe.

Keep review and accountability visible

Specify what the system may support, what remains a human decision, and who checks outputs before they inform advice or reach a client. The reviewer should be able to inspect relevant source material, test uncertain statements, correct the output and decide not to use it.

Keep a proportionate decision record covering the task, information involved, provider evidence, applicable controls, accountable owner, review point and conditions that trigger reassessment. If the firm cannot establish a key part of the information path or review arrangement, pause that proposed use of the material while the evidence gap is addressed.

Choose a practical next step

Pattrn Data’s AI Risk & Efficiency Audit can help a firm map AI uses, information flows, owners, hand-offs, review points and controls before wider rollout or tool spend. For a focused question, an AI Clarity Consultation can help frame a bounded next step. Neither replaces qualified legal advice on a specific matter.

Source and scope

This article was informed by the Law Society Journal discussion of client generative-AI use and privilege risk. That discussion includes examples from multiple jurisdictions and is not a statement of law for every firm or matter. The checklist above is general operational guidance, not legal advice.

Frequently asked questions

Does using an AI tool automatically waive privilege?

This article does not make that legal conclusion. The answer depends on the circumstances and applicable law. A firm should seek qualified legal advice where a matter-specific decision is required.

Is a provider’s general privacy statement enough?

Not by itself for this assessment. The firm should gather evidence about the specific service, settings, information flows, access, retention and relevant contractual terms, then compare it with the client and matter requirements.

What should the firm do if it cannot establish where material goes?

Record the uncertainty and pause the proposed use of that material until the information path and relevant controls can be assessed.

Discuss your next step

Talk through your situation with Rohit, or take the assessment if you need a first read.