Law Firm Data Breach Assessment Notes: Use case and AI Confidentiality

When a client’s mailbox is compromised, the law firm is asked the same day whether the notifiable data breaches scheme applies. The answer is written up as a file note: what happened, what was taken, whether serious harm is likely and when the deadlines fall. It is written under pressure, so staff paste the facts into ChatGPT or Claude to structure it. This page marks every item in an assessment note for a property agency that rule 9, privilege and the Privacy Act bear on, explains what each one exposes, and shows how the same note is built on Ermos inside the firm.

The data breach assessment note

Written within days of an incident for the client and the partner, updated as the forensic report arrives and kept on the matter file.

File note: is this an eligible data breach?
Client: · matter TIDE-0431 · Tuesday 10 November 2026 ·

1. What happened.

2. What was in the mailbox.

3. Is Tidewater covered?

4. Timing.

5. Serious harm.

6. Remedial action.

7. Recommendation.

8. Client instructions.

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

Breach assessment note drafter

An Agent+ that prepares and updates a data breach assessment note for a client. It reads the incident report and forensic findings as they arrive, lists the data types involved, and sets the facts against your own checklist. Unknowns and changes since the last version are marked.

What it does

  1. Reads the incident report and forensic updates
  2. Organises facts against your own assessment checklist
  3. Marks unknowns and what changed since the last version

What stays with your people. A partner forms the assessment and advises the client. The agent organises facts; it does not decide whether the breach is notifiable.

An Agent+ is built once and anyone in the practice can run it. It works from your own documents and runs on Ermos Dominion, the AI unit in your office, or on Ermos Edge, your own instance in an Australian data centre.

What leaves the firm

  1. 1
    The client and the matter

    A named business, its principal and the fact that it has had a security incident serious enough to need legal advice. Confidential to the client under rule 9, and reputationally sensitive until the client chooses to notify. Rule 9 confidentiality guide.

  2. 2
    The incident narrative and the staff member

    How the compromise happened, the dates and a named employee who entered her password on a fake page. Personal information about a staff member who made a mistake, and the client’s confidential account of its own failure. Privacy Act guide.

  3. 3
    What was in the mailbox

    The number of tenancy applications, the kinds of document and the number of landlords whose bank details were copied. No applicant is named, but the paragraph states how bad the breach is for people the client holds information about. Rule 9 confidentiality guide.

  4. 4
    Coverage and the client’s turnover

    The client’s turnover against the $3,000,000 small business test in section 6D of the Privacy Act, which asks about annual turnover for the previous financial year, with exceptions in section 6D(4). A confidential financial figure, and the basis on which the Act applies to the client. Privacy Act guide.

  5. 5
    The 30 day assessment rule and the client’s dates

    Section 26WH requires a reasonable and expeditious assessment and all reasonable steps to complete it within 30 days after the entity becomes aware. The dates tie the rule to a named client’s incident. Privacy Act guide.

  6. 6
    The serious harm assessment

    The firm’s reasoning against the factors in section 26WG, applied to the client’s facts. The central privileged judgement in the note, and the one a regulator or a class-action lawyer would most like to read. Legal privilege guide.

  7. 7
    The remedial action analysis

    Why the remedial action exception in section 26WF is not available on current facts and what the forensic report may change. Privileged advice on whether the client has to notify at all. Legal privilege guide.

  8. 8
    The recommendation and the statement

    Section 26WK sets out what the statement to the Commissioner must contain, and section 26WL requires the entity to notify the individuals, or those at risk, as soon as practicable after preparing it. The recommendation is the firm’s advice to the client on a decision it has not yet taken. Legal privilege guide.

  9. 9
    The client’s wish and the privilege strategy

    What the client asked the firm to say, the firm’s refusal, and the firm’s own doubts about its privilege claim over the forensic report. Candid internal advice that would damage the client if disclosed. Legal privilege guide.

  10. 10
    The solicitors

    Initials that tie the note to practitioners bound by rule 9 and supervised by a partner. 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 Use cases in law firms.

Common questions

Do ChatGPT and Claude use the information in a data breach assessment note?
Everything pasted or uploaded is sent to the AI company's servers and processed there to produce a reply. 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. This note carries Daniela Rusu's password entered on a fake login page, the 212 tenancy applications and 38 landlords' bank details in the copied mailbox, and the firm's view that serious harm is likely.
Is it safe to paste a data breach assessment note into ChatGPT or Claude?
No, because the note holds the firm's privileged judgement on whether the client must notify, and its doubts about its own privilege claim over the forensic report. 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. Leaving out Tidewater Realty still leaves the dates of 16, 19 and 28 October, 212 applications and a turnover of $4.31 million, which point to one property agency. 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.
Can ChatGPT or Claude remember the details in a data breach assessment note later?
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). For this note, the details that could carry forward are the client's account of its own failure, the number of people affected and the client's request to say there was no evidence of misuse, which the firm advised against.
How can a law firm use AI for a data breach assessment note 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 note is drafted from the incident facts, the forensic report and the firm's own file note template. A solicitor decides whether serious harm is likely and whether the client must notify.
What is an eligible data breach under the Privacy Act?
Under section 26WE, it is unauthorised access to or disclosure of personal information held by an APP entity, or loss of that information where unauthorised access or disclosure is likely, in each case where a reasonable person would conclude serious harm to any of the individuals is likely. Section 26WF removes cases where timely remedial action means serious harm is no longer likely.
How long does an organisation have to assess a suspected data breach?
Section 26WH requires an entity that is aware of reasonable grounds to suspect an eligible data breach to carry out a reasonable and expeditious assessment and take all reasonable steps to complete it within 30 days after becoming aware. Once it has reasonable grounds to believe there has been an eligible data breach, it must prepare a statement and give it to the Commissioner as soon as practicable (section 26WK).
What must a data breach statement contain?
Under section 26WK(3), the statement must set out the entity's identity and contact details, a description of the breach, the kinds of information concerned and recommendations about the steps individuals should take. The entity must then notify the individuals concerned, or those at risk, as soon as practicable after preparing it (section 26WL), or, if neither is practicable, publish the statement on its website and take reasonable steps to publicise it.

See Dominion answering from your own documents.

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

Book a demo