Care plans and AI confidentiality: what a resident’s plan gives away in ChatGPT or Claude.

The care plan is the most complete record a provider holds about a resident: diagnoses, medications, how they move, how they behave, what they fear and who makes decisions for them. It is also the document most often pasted into ChatGPT or Claude, usually to be summarised for a new staff member or rewritten in plainer language for a family. This page marks every item in a routine care plan that the Aged Care Act 2024 and the Privacy Act protect, which is nearly all of it, and shows how the same plan is summarised on Ermos inside the service.

The resident care plan

Prepared on admission and reviewed at set intervals or after a change in condition, by the registered nurse with the resident, family and GP.

Kurrajong Gardens Aged Care
Care plan summary · , · · Reviewed 18 October 2026

Diagnoses.

Medications.

Mobility and falls.

Cognition and behaviour.

Continence and skin.

Decision-making.

Social.

An invented care plan. The Medicare number is a published test value. Hover or tap any highlighted item for the analysis; the full list is below.

What leaves the service

  1. 1
    The resident and her room

    A named person and where she sleeps. Personal information the Act requires the provider to protect, and a location. Privacy Act guide.

  2. 2
    Date of birth, My Aged Care ID and Medicare number

    Three identifiers used across government and health systems. Together with a name, the material identity fraud is built from, about a person who cannot monitor her own accounts. Privacy Act guide.

  3. 3
    Diagnoses

    Health information, which is sensitive information under the Privacy Act. Section 168 of the Aged Care Act permits its use only for purposes connected with delivering her care. Patient data guide.

  4. 4
    Medications, including a psychotropic

    A full medication list, including risperidone, whose use for behaviour is a restrictive practice under the Act and subject to its own authorisation rules. Patient data guide.

  5. 5
    Falls history and risk

    Clinical information about injury and risk, which also bears on the provider’s own incident and quality obligations. Patient data guide.

  6. 6
    Behaviour, distress and a late husband

    The most intimate paragraph in the plan: how a woman with dementia suffers in the evenings and what comforts her. Written for the people who care for her, and now on a third party’s servers. Patient data guide.

  7. 7
    Continence and a pressure injury

    Personal care details and a wound. Health information, and the kind of detail that would humiliate her if she could read it anywhere else. Patient data guide.

  8. 8
    The decision-makers and the advance care directive

    Two family members, a phone number, and her end-of-life wishes. Three people’s personal information and the most consequential instruction in the file. Privacy Act guide.

  9. 9
    A family estrangement

    A confidence about her family and an instruction about who may be told what. Information given to the provider for one purpose only. Privacy Act 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 service 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 resident’s record. 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 aged care providers.

Common questions

Do ChatGPT and Claude use the information in a resident care plan?
Yes, a pasted care plan is sent to the AI company and processed there to produce the 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 plan puts the diagnoses, the psychotropic and the advance care directive into a single paste.
Can we use ChatGPT or Claude for a care plan if we remove the resident's name?
No, because a care plan without a name still describes an identifiable person. 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 date of birth, a hip fracture repaired in 2024 and a late husband the resident calls for in the evenings point to one person. The Medicare number identifies that person outright. 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.
Is it safe to use an AI note-taker or meeting recorder for a resident care plan review?
Only once the service knows who receives the recording. Recording and transcription tools send the audio or transcript to the tool's vendor. The same questions apply as for ChatGPT and Claude (where it is stored, whether it trains a model, who can review it). The OAIC's guidance also tells organisations to review carefully the terms and settings that will apply to their use of any AI product. A care plan review brings the resident, family and GP together, so the transcript would carry the diagnoses, the psychotropic and any family estrangement in their own words.
How can an aged care service use AI for a resident care plan 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. A staff member asks for a summary or a plain-language version. It is built from the resident's record in the service's own care plan format, and a registered nurse checks and signs it.
Does the Aged Care Act 2024 stop us using AI with care plans?
The Act does not mention AI. Section 168 requires a registered provider to ensure the protection of personal information and limits its use and disclosure to purposes connected with delivering the person's care, except with consent. Whether transmitting a care plan to a frontier model provider is such a purpose, and whether the resident or their decision-maker has consented, are the questions a provider has to be able to answer. Processing inside the service makes both questions short.
Staff paste care plans to summarise them for handover. Is that the problem?
The use is sensible; the destination is the problem. On an Ermos unit the plan is summarised on hardware in your own building, and nothing in it leaves. On Edge the record stays on your own instance in Australia and only permitted excerpts go, per question, to a named Australian provider under contract.
What about the resident's capacity to consent?
A resident with moderate dementia may not be able to consent to a new use of her information, which is why the Act provides for supporters and substitute decision-makers. In practice it means nobody has consented to the care plan going to a frontier model, and nobody is in a position to.

See Dominion answering from your own documents.

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

Book a demo
Keep reading
AI data leakage in aged care providers AI for healthcare AI and patient data All professions