GP Ahpra Mandatory Notification Reports: Use case and AI Confidentiality

A doctor who has watched another doctor practise unsafely has to decide, quickly and often on little sleep, what the law requires and how to say it. An Ahpra mandatory notification report names the colleague, sets out what the notifier saw and usually describes patients who were affected. People drafting one worry about the wording, so they paste their notes into ChatGPT or Claude and ask for a neutral, structured report. This page marks every item in an invented draft notification that the National Law, 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 Ahpra mandatory notification report

Written by a registered health practitioner who has formed a reasonable belief that another practitioner has engaged in notifiable conduct, then lodged with Ahpra through its online portal. The draft often passes through the practice manager and the doctor’s insurer first.

Wombeyan Street Family Practice
8 Wombeyan Street, Eastwood NSW 2122 · (02) 5550 0161

Mandatory notification to Ahpra: draft for the practice file (not yet lodged) · Tuesday 10 November 2026

Notifier.

Practitioner notified.

Type of notifiable conduct.

Basis for my belief.

Events.

Witness.

Action taken.

Drafting notes, kept off the notification.

An invented draft notification. The practice, doctors, nurse, patients and hospital are fictional, the registration numbers are placeholders and the mobile number is masked. Hover or tap any highlighted item for the analysis; the full list is below.

Notification report drafter

An Agent+ that helps a practitioner write up a concern about a colleague. It works from the incident notes, statements and timeline the practice already holds, and produces a draft in your template that keeps facts apart from opinion and marks anything missing.

What it does

  1. Reads the incident notes, statements and timeline
  2. Drafts the report in your template, facts separated from opinion
  3. Lists gaps and unsupported statements for the practitioner

What stays with your people. The practitioner who formed the belief decides whether the threshold is met and makes any report. The practice manager and insurer may read the draft first.

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 practice

  1. 1
    The notifier and her request for confidentiality

    A register name, a masked registration number and a mobile, with a request that her identity is kept confidential. Ahpra’s page says it will try to keep a mandatory notifier’s identity confidential if asked. Once the draft is in a vendor’s system, that request depends on a different organisation.

  2. 2
    The colleague and his engagement dates

    Personal information about the doctor the notification is about, and the dates he worked at the practice. Section 141 is about the conduct of another registered health practitioner. A draft that names him is an allegation about a named person that he has not yet seen or answered.

  3. 3
    The kinds of notifiable conduct

    Section 140 lists four kinds: practising while intoxicated by alcohol or drugs, sexual misconduct, placing the public at risk of substantial harm because of an impairment, and placing the public at risk of harm through a significant departure from accepted professional standards. The words are public. Choosing which kind fits is the notifier’s judgment.

  4. 4
    What she saw and what she was told

    Section 141(1) turns on a reasonable belief formed in the course of practising. Ahpra’s page gives direct knowledge, or a report from a reliable source, as the example of a sound basis, and says a notification should not be based on rumours or gossip. Keeping first-hand and second-hand apart is the point of this paragraph, and a tidy-up can blur it.

  5. 5
    A named patient, his symptoms and his heart attack

    A patient, his age, a presentation, what was and was not done, and a hospital admission. Health information is sensitive information (Privacy Act s 6(1)) and covers information collected in providing a health service (s 6FA). APP 6.2(b) allows a disclosure that is required or authorised by or under an Australian law, which bears on what goes to Ahpra. It does not extend to an AI provider. Patient data guide.

  6. 6
    A child and a dose error

    A six-year-old, a medicine, an error ten times too large and the pharmacy that caught it. Health information about a child, held by the practice. His mother is not named here and the pharmacy’s name ties the event to a local business. Patient data guide.

  7. 7
    The nurse and her willingness to speak

    Personal information about a named colleague and what she observed. She has agreed to speak to Ahpra, which is her decision, and not to a software vendor.

  8. 8
    The practice’s own actions

    That the locum was told not to attend and the engagement ended, and that the insurer has been told. Ahpra’s page says it may ask what has been done to manage the risk, so this is part of the story, and it is also a record of the practice’s response.

  9. 9
    A rumour left out on purpose

    Hearsay about a colleague’s drinking at a previous practice. The doctor has left it out of the notification, which is consistent with Ahpra’s page on rumours and gossip. It still sits in the file as an unverified statement about a named person, and opinions about someone’s health can be personal information that APP 11.1 asks the practice to take reasonable steps to protect.

  10. 10
    A colleague’s plea and a timing question

    Section 141(2) says the practitioner must notify as soon as practicable after forming the reasonable belief. Section 141(3) says a contravention is not an offence but may be behaviour for which action may be taken. Section 141(4) sets out situations in which the belief is treated as not formed in the course of practising, including knowing or reasonably believing that Ahpra has already been notified (paragraph (e)); a colleague’s request is not among them. Good medical practice 8.3.4 and 8.3.5 ask doctors to help a colleague to get help and to comply with statutory reporting requirements.

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 Use cases in general practitioner (GP) practices.

Common questions

Do ChatGPT and Claude use the information in an Ahpra mandatory notification report?
Yes, a pasted draft notification 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 draft names a locum doctor with his registration number and engagement dates, a 58-year-old with chest pain, a six-year-old's antibiotic dose, a nurse witness and the notifier's own mobile number and request to keep her identity confidential.
Is it safe to paste a draft Ahpra mandatory notification into ChatGPT or Claude?
No, because the draft is an allegation about a named doctor that also holds patient health information. 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. Taking the names out would not help much. 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 locum role, the dates, the hospital and the pharmacy still point to one practice and one doctor who has not yet seen or answered the allegation.
What if a doctor uses her own ChatGPT or Claude account for a mandatory notification draft?
The practice cannot see or manage what was sent. 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. Both companies can also review content that is flagged for safety or policy reasons. For this draft that means a colleague's name, two patients and a nurse's account could sit in a personal account the practice does not control. Ahpra says it will try to keep a notifier's identity confidential if asked, and once the draft is in a vendor's system that request depends on a different organisation.
How can a GP practice use AI for a mandatory notification report 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 draft is produced from the doctor's own notes and the practice's template. She reads every line, keeps what she saw apart from what she was told, and takes the timing question to her insurer or adviser before anything is lodged.
Who has to make a mandatory notification about another doctor?
Under section 141 of the National Law, a registered health practitioner who, in the course of practising, forms a reasonable belief that another registered practitioner has behaved in a way that constitutes notifiable conduct must notify the National Agency (Ahpra) as soon as practicable. Employers have a separate duty under section 142 and treating practitioners are dealt with in sections 141A and 141B. Section 141(4) lists situations in which the belief is treated as not formed in the course of practising, including knowing or reasonably believing that Ahpra has already been notified.
What counts as notifiable conduct?
Section 140 lists four kinds: practising while intoxicated by alcohol or drugs, sexual misconduct connected with practice, placing the public at risk of substantial harm because of an impairment, and placing the public at risk of harm through a significant departure from accepted professional standards. Ahpra says a notification should not be based on rumours or gossip, and gives direct knowledge, or a report from a reliable source, as the example of a sound basis.
Is a doctor who notifies in good faith protected?
Section 237 says a person who in good faith makes a notification is not liable civilly, criminally or under an administrative process for giving the information. The notification is not a breach of professional etiquette or ethics and no liability for defamation is incurred. Ahpra says it will try to keep a mandatory notifier's identity confidential if asked.

See Dominion answering from your own documents.

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

Book a demo