Briefs to counsel and AI data leakage: what a whole matter gives away in ChatGPT or Claude.

A brief to counsel is the matter, condensed: a chronology of everything, the evidence for and against, the weaknesses the firm has found, and the questions it wants answered. It is the longest document a solicitor writes and the one most tempting to have ChatGPT or Claude summarise or structure. It is also the maximum exposure a single document can carry. This page marks every item in a brief’s observations that rule 9 and privilege protect, and shows how the same brief is assembled on Ermos inside the firm.

The brief to counsel

Prepared before counsel is briefed to advise or appear, with the documents indexed and the solicitor’s observations attached.

Brief to counsel: observations
· To: · From:

Chronology (summary).

Evidence for.

Weaknesses.

Medical.

Questions for counsel.

An invented brief. Hover or tap any highlighted item for the analysis; the full list is below.

What leaves the firm

  1. 1
    The matter

    Named parties. The whole brief is identifiable. Rule 9 confidentiality guide.

  2. 2
    Counsel

    A named barrister, briefed in a matter. A third party’s personal information and a professional relationship. Privacy Act guide.

  3. 3
    The chronology

    The entire matter in ten sentences, including document numbers from the other side’s discovery, which are subject to the obligation in Hearne v Street. Legal privilege guide.

  4. 4
    The evidence

    What the firm will rely on. Litigation strategy, privileged. Legal privilege guide.

  5. 5
    Every weakness

    The firm’s candid assessment of its own case, including a question about the client’s own conduct. The single most privileged paragraph in any matter, and the one the other side could win with. Legal privilege guide.

  6. 6
    The client’s health and a request

    Sensitive health information, a treating doctor, and a request about how it is handled in court. Protected under the Privacy Act and by rule 9. Privacy Act guide.

  7. 7
    The questions, and the client’s bottom line

    What the firm wants to know, and the figure the client will accept. The settlement authority, written into the brief. Legal privilege guide.

  8. 8
    Counsel’s fees

    Billing arrangements between the firm, the client and counsel. Confidential, and under the Law Society guidance a matter of care. Law Society AI guidance.

  9. 9
    The solicitors

    Ties the brief to the practitioners responsible for it. Supervising AI guide.

What happens in Frontier Models (ChatGPT, Claude)

  • The whole document is transmitted to the provider’s servers. On personal plans the user cannot choose where it is stored or processed, and for the major providers that has generally meant outside Australia, under that country’s legal process. Some business plans offer storage in Australia, but the provider still receives the content.
  • On consumer tiers, prompts can be retained and used to train future models.
  • The provider’s staff and systems can reach the content under the provider’s own policies.
  • Once sent, the firm no longer controls it. Retention and deletion happen on the provider’s terms.

The same document on Ermos

Ermos Dominion

The document is drafted on a unit in your office, from your own templates and the matter file. Every marked item stays on the unit, which keeps working with the internet switched off. Nothing in it is transmitted to an AI provider, so there is no disclosure to analyse: the confidentiality work is the same as for your file server.

Ermos Edge

Your documents, the index and your access boundaries stay on your own instance in an Australian data centre. When a staff member asks for this draft, only their question and the excerpts they are permitted to see go to a named Australian provider, who is contractually barred from keeping the content or training any model on it. Nothing is processed outside Australia.

Frontier Models (ChatGPT, Claude)Ermos EdgeErmos Dominion
Where the document is processedProvider servers, generally offshore on personal plansYour instance in an Australian data centre; answers generated in AustraliaThe unit in your office
Who can read itProvider staff and systems, under policyYour staff, within the boundaries you set. The named Australian provider processes only permitted excerpts, under contract, and keeps nothingYour staff only
Is it keptOn the provider’s termsKept on your instance; the provider is barred from retaining itStays on the unit
Can it train a modelOn consumer tiers, yesNo, by contractOnly your own unit learns from it, on your hardware

Part of AI data leakage in law firms. See how this works in practice: Chronologies on Ermos.

Common questions

Do ChatGPT and Claude use the information in a brief to counsel?
Yes, a pasted brief is sent to the AI company's servers to be processed. As at October 2026, whether a conversation on a personal plan (ChatGPT Free, Plus or Pro; Claude Free, Pro or Max) can be used to train the models depends on a setting each user controls. Business plans (ChatGPT Business and Enterprise, Claude Team and Enterprise) and the developer APIs sit under commercial terms, and OpenAI and Anthropic each say they do not train on that content by default (OpenAI, Anthropic). On either product, the thumbs up and thumbs down buttons can send a whole conversation back for training. A brief to counsel holds the chronology, every weakness the firm has found and the client's settlement authority.
Is it safe to paste a brief to counsel into ChatGPT or Claude if we take the names out?
No, because a brief to counsel stays identifiable after the names go; the chronology alone describes the whole case. Removing names rarely makes a document anonymous. Information stays personal information while the person is reasonably identifiable, and the OAIC warns that information can be at risk of re-identification even when it has been de-identified or anonymised. The discovered document numbers, the medical report and the settlement authority point to one matter. The Office of the Australian Information Commissioner (OAIC) recommends in its guidance on commercially available AI products that organisations do not enter personal information, and particularly sensitive information, into publicly available generative AI tools, because of the significant and complex privacy risks.
Can ChatGPT or Claude remember the details in a brief to counsel later?
They can. Both products also have memory features that can carry details from one conversation into later ones, and on a personal account each user decides whether they are switched on (ChatGPT, Claude). A brief to counsel holds the client's weaknesses, her medical history, the discovered document numbers and the figure she will accept. A solicitor who works on several matters in one account would need to check which details the memory feature has carried over before starting the next matter.
How can a law firm use AI for a brief to counsel without sending it to ChatGPT or Claude?
Ermos is built so the document stays under your control. On the Dominion AI unit it is drafted inside your office and none of it is sent to an AI provider. On Ermos Edge your documents and index stay on your own instance in an Australian data centre, and only the question and the excerpts the user is permitted to see go to a named Australian provider that is contractually barred from keeping them or training any model on them. The solicitor asks for the chronology in plain language, and it is built from the indexed matter file and the firm's brief template. Each event is traceable to its source, and the solicitor checks every entry before signing the brief.
Why is the brief worse than the advice letter?
The advice letter is what the firm decided to tell the client. The brief is what the firm actually thinks, including the weaknesses and the client's conduct, written for a barrister and nobody else. It is the most candid document in the matter and the longest, which is why it is the one staff most want summarised.
Can AI help assemble a brief without the matter leaving the firm?
On Ermos the chronology is built from the documents on the unit in your office, each event traceable to its source, and the observations are drafted from the file. Counsel still receives the brief the usual way. Nothing in it has been anywhere else.
What about the discovered document numbers in the chronology?
They tie the brief to documents produced by the other side under compulsion, which the obligation restated in Hearne v Street restricts to the proceeding. A brief to counsel is squarely within the proceeding. A frontier model provider's servers are not.

See Dominion answering from your own documents.

A demonstration takes 30 minutes, uses no client data, and comes with no obligation.

Book a demo