Participant support plans and AI confidentiality: what an NDIS support plan gives away in ChatGPT or Claude.

The participant support plan tells a new worker who the participant is, how to support him and what to do when something goes wrong. It is also the document a team leader pastes into ChatGPT or Claude to shorten for a casual worker, turn into Easy Read for the participant, or tidy before a plan review. It holds a diagnosis, medication, triggers, the way into the house, family arrangements and things the participant has asked to keep private. This page marks every item in a routine NDIS support plan that the Practice Standards, the Code of Conduct and the Privacy Act bear on, and shows how the same plan is summarised on Ermos inside the service.

The participant support plan

Written by the team leader with the participant and his nominee when supports start, and reviewed at each plan review or after a change. The Support Planning standard expects the participant to be involved and the plan to be understandable and accessible to him and to workers.

Brushbox Community Supports
Participant support plan · · Reviewed 12 October 2026

Health
Plan and goals
What overwhelms him
Home
Family
Private
Supports
Sharing
Emergency

An invented support plan. The NDIS number and phone number are placeholders. Hover or tap any highlighted item for the analysis; the full list is below.

What leaves the service

  1. 1
    The participant, address and NDIS number

    A named man, where he lives alone and the number the NDIA uses to match payments to him. Personal information, and the material that doorstep scams and claims fraud are built from. The Information Management standard expects each participant’s information to be identifiable, accurately recorded, current and confidential. Privacy Act guide.

  2. 2
    Diagnoses and medication

    Autism, an intellectual disability, an anxiety diagnosis and the dose of his medication. Health information, which is sensitive information under the Privacy Act. The OAIC lists disability service providers that handle health information among the organisations the Act covers, whatever their size. Patient data guide.

  3. 3
    Plan dates and goals

    The period of his NDIS plan and what he wants for his life. Personal, and the Privacy and Dignity standard expects the participant to understand and agree to what personal information is collected and why. Privacy Act guide.

  4. 4
    Triggers and what settles him

    How he reacts to crowds and touch, that he hits his own temple when overwhelmed, and what helps. Written so any worker can support him safely, and the most intimate paragraph in the plan. Health information, and the kind of detail that could be used to characterise him unfairly. Patient data guide.

  5. 5
    Entry arrangements

    That he lives alone, where the spare key is kept and who holds the code. Bears on his physical safety as much as his privacy, and on the Code of Conduct duty to provide supports in a safe and competent manner.

  6. 6
    The nominee and a request about his father

    His mother’s name and number, and an instruction that his father receive no information. Two other people’s personal information and a rule about who may be told what. Privacy Act guide.

  7. 7
    A relationship he asked to keep private

    That he has a girlfriend and has asked that his mother not be told. A confidence given for one purpose. The Code of Conduct requires respect for the privacy of people with disability and for their right to self-determination and decision-making. Privacy Act guide.

  8. 8
    Who supports him, and when

    Worker names and the days and hours he has support. A map of when a man who lives alone is, and is not, with a worker, and personal information about the workers. Privacy Act guide.

  9. 9
    The record of who the plan may be shared with

    The Support Planning standard has plans shared, with consent, with the support network, other providers and relevant government agencies. An AI company is not on Daniel’s list, and the Information Management standard expects consent to cover the purpose and when information may be disclosed. Privacy Act guide.

  10. 10
    Emergency steps

    Generic wording that appears in every plan the service writes. It names no one.

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 participant 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 NDIS services.

Common questions

Do ChatGPT and Claude use the information in a participant support plan?
Yes. Everything pasted into ChatGPT or Claude 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. A plan pasted in to make an Easy Read copy carries Daniel's diagnoses and sertraline dose, the way he reacts to crowds and a relationship he has asked his mother not be told about.
Can we use ChatGPT or Claude for a support plan if we remove the names?
Not safely, because the plan still describes one person without them. 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. A man who lives alone, aims to catch the train to TAFE without a worker, takes sertraline each morning and has a nominee mother and an excluded father is recognisable to anyone who knows him. 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 participant support plan 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 a support plan that matters because Daniel's triggers, the place his spare key is and his instruction about his father are details a team leader would not want carried into any later conversation. Each plan pasted adds another participant to the same account.
How can an NDIS provider use AI for a participant support 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. The summary or Easy Read copy is produced from the participant file and the service's own plan template. The team leader checks and signs it before a worker relies on it, and the participant's decisions about who may see the plan still apply.
Does a participant support plan count as health information?
In most cases it does. A support plan usually records a disability, medication and how a person is supported, which is information about their health. The OAIC says an organisation that provides a health service and holds health information is covered by the Privacy Act, and lists disability service providers that handle health information as an example. Whether a particular service is covered depends on the services it delivers.
What do the NDIS Practice Standards say about who can see a support plan?
The Support Planning standard says support plans are given to participants in a form they can understand and, with consent, shared with the support network, other providers and relevant government agencies. The Information Management standard expects consent to collect, use, retain or disclose information, including the purpose and when disclosure may occur without consent. Neither mentions AI.
Can staff use AI to make an Easy Read version of a support plan?
Easy Read versions help participants understand their plans, and the Practice Standards expect plans to be understandable. The question is where the conversion happens. The Commission's February 2026 position statement on behaviour support plans expects participant information to be de-identified if AI is used. On an Ermos unit the plan is converted on hardware in your own building and nothing in it leaves. Whether that meets the Commission's expectation is for the service to decide, with its own advice.

See Dominion answering from your own documents.

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

Book a demo