Insurer medical reports and AI patient privacy: what a GP report to an insurer gives away in ChatGPT or Claude.

A GP report to an insurer is the most detailed account of a patient the practice writes for anyone outside it. An income protection insurer asks for the diagnosis, the treatment, the prior history, the lifestyle and the doctor’s opinion about work capacity, and the patient has signed an authority for exactly that. The report is long and repetitive, so it is pasted into ChatGPT or Claude to draft or shorten. This page marks every item in a routine insurer report that the Privacy Act, Good medical practice and the NSW health records law bear on, and shows how the same report is drafted on Ermos inside the practice.

The medical report to an insurer

Written when a patient claims on an income protection or life policy and signs an authority for the insurer to obtain the GP’s report. The authority sets how much the doctor may say.

Kestrel Lane Medical Practice
Medical report for Alderpoint Life · 28 October 2026

Re:

Authority.

Occupation.

Diagnosis.

Treatment.

Prior history.

Lifestyle.

Opinion.

Finances.

An invented report. The insurer and claim number are fictional and the registration number is a placeholder. Hover or tap any highlighted item for the analysis; the full list is below.

What leaves the practice

  1. 1
    The patient and the claim reference

    Name, date of birth and an insurer’s claim number. A live claim, tied to a person. Privacy Act guide.

  2. 2
    Occupation and the physical demands of the work

    A named business and what the work involves. Employment information, and the factual base for the opinion on capacity. Privacy Act guide.

  3. 3
    Diagnosis, imaging and the cause

    A diagnosis, the scan and where it was done, and how the injury happened. Health information is sensitive information under the Privacy Act. Patient data guide.

  4. 4
    Treatment, medication and the specialist

    Named physiotherapist and surgeon, medications and a date. Health information and a map of everyone involved in the patient’s care. Patient data guide.

  5. 5
    Prior back history and past depression

    Earlier episodes the insurer asked about, including a mental health history from 2023. Disclosed to an insurer under authority, and a different matter in a chat tool. Sensitive health information. Patient data guide.

  6. 6
    Smoking, alcohol and BMI

    Lifestyle facts that an insurer weighs. Health information given to the doctor in confidence. Patient data guide.

  7. 7
    The doctor’s opinion on capacity

    The GP’s view on fitness for work and the expected return date. Clinical and professional judgement that a doctor must form and own.

  8. 8
    Income, mortgage and fear of losing the house

    Financial detail the patient shared to explain his situation. Personal information, and context the insurer did not ask for. Privacy Act guide.

  9. 9
    The scope of the patient’s authority

    The authority sets what the doctor may disclose to the insurer. APP 6 turns on consent and on what a patient would reasonably expect, and it names the insurer only, leaving an AI tool out of it. Privacy Act guide.

  10. 10
    The doctor and registration number

    Ties the report to a named doctor bound by Good medical practice. Registration details are public, but they identify the source.

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 practice 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 patient 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 healthcare practices. See how this works in practice: Correspondence summaries on Ermos.

Common questions

Do ChatGPT and Claude use the information in a medical report to an insurer?
Yes, a pasted insurer report 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 report holds a claim number, a lumbar diagnosis, past depression from 2023, about 20 drinks a week, an opinion on work capacity, and the patient's income and mortgage.
Is it safe to paste a medical report to an insurer into ChatGPT or Claude if we remove the patient's name?
No, because the clinical detail still points to him. 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 self-employed electrician with an L4/5 disc injury on 3 August 2026, a named physiotherapist and a named neurosurgeon is easy to recognise. His signed authority names the insurer, so an AI tool sits outside it. 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.
Does a ChatGPT Business or Claude Team plan make it safe for a medical report to an insurer?
Not on its own. A business plan changes the default on training, but it does not change who receives the content. On a personal plan the user cannot choose where conversations are stored or processed. Some business plans let an organisation choose to store data at rest in Australia (OpenAI), which still leaves the content with the AI company. The OAIC's guidance also tells organisations to review carefully the terms and settings that will apply to their use of any AI product. The patient signed an authority for the insurer, and a report sent to an AI company reaches a recipient that authority does not name. Our comparisons with ChatGPT Enterprise and Claude weigh the trade-offs.
How can a medical practice use AI for a medical report to an insurer 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 report is drafted from the patient record and the practice's own insurer report template. The GP checks it against the authority, forms the opinion on fitness for work and signs it.
The patient signed an authority for the insurer. Does that cover an AI tool?
The authority covers the insurer for the claim. APP 6.1 prohibits using or disclosing information for another purpose without consent or an APP 6.2 exception, and under the reasonable-expectation exception a secondary purpose for sensitive information must be directly related to the primary one. A third-party AI provider is a further recipient the patient did not name.
Does an overseas AI provider add anything?
Yes. APP 8.1 requires an entity, before it discloses personal information to an overseas recipient, to take reasonable steps to ensure the recipient does not breach the APPs, subject to the exceptions in APP 8.2. For a report on a patient's diagnosis, prior depression and finances, that is a significant step to take before pasting.
Who forms the opinion on fitness for work?
The treating doctor. Good medical practice expects clear, accurate records and professional judgement, and Ahpra says practitioners must apply human judgment to any AI output. A draft can be shorter or better organised, but the opinion that goes to the insurer is the doctor's own.

See Dominion answering from your own documents.

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

Book a demo