Dental radiology reports and patient privacy: what a CBCT or X-ray report gives away in ChatGPT or Claude.

A dental radiology report comes back from the imaging centre in technical language, and the dentist or a nurse wants it in plain English, either to explain to the patient or to paste into the chart. That is a common task for ChatGPT or Claude. The report carries the patient’s identity, the clinical question, findings about nerves and bone, often an incidental finding unrelated to teeth, and a reference to the images themselves. This page marks every item in a routine CBCT report that the Privacy Act and the Dental Board’s Code of conduct bear on, and shows how the same report is summarised on Ermos inside the practice.

The dental radiology report

Returned by an imaging centre after a scan requested by the dentist. Section 8.3 of the Code of conduct lists investigations among the things a health record should cover, and asks that the record be held securely.

Cone beam CT report, limited volume
· 21 October 2026

Patient
Clinical question
Technique
Findings: lower left
Findings: lower right
Incidental
Impression
Images

An invented radiology report. The imaging centre, radiologist, dentists and patient are fictional; the accession 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 referrer

    A name and date of birth beside a named referring dentist. Personal information under the Privacy Act and health information as it is collected in providing a health service. Patient data guide.

  2. 2
    The clinical question

    Why the scan was requested: a recurring infection and an upcoming surgery under sedation. Health information, and a record of the dentist’s reasoning. Patient data guide.

  3. 3
    Technique

    Scan parameters and the reconstructions viewed. Generic technical wording, with no information about the patient other than a date.

  4. 4
    Findings about the teeth and the nerve

    Anatomical relationships for named teeth, including contact with the nerve canal. Clinical content the Code of conduct (section 8.3) expects the record to hold, and health information under the Privacy Act. Patient data guide.

  5. 5
    An incidental finding in the neck

    A possible vascular calcification unrelated to teeth, with a recommendation for medical review. Sensitive information the patient may not yet have been told, and a result that a person must read carefully rather than a summary. Patient data guide.

  6. 6
    A sinus finding

    A small mucosal thickening noted in passing. Health information, usually minor, but still about a named patient. Patient data guide.

  7. 7
    The impression and recommendations

    The radiologist’s conclusion and what to do next, including a risk discussion with the patient. The summary the dentist will rely on, and the patient’s consent conversation depends on it (Code, section 4.2).

  8. 8
    The imaging centre and radiologist

    A named radiologist and practice. Public details of a registered practitioner, tied to the scan.

  9. 9
    The accession number and image access

    The reference that unlocks the images for 90 days. An identifier for the scan; anyone with access to the viewer could use it. 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 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 dental practices.

Common questions

Do ChatGPT and Claude use the information in a dental radiology report?
Yes, a pasted radiology 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 links a patient's wisdom teeth to the nerve canal, notes a possible carotid calcification and gives an accession number that opens the images for 90 days.
Is it safe to paste a dental radiology report into ChatGPT or Claude if we remove the patient's name?
No, because the report still points to one patient. 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 date of birth, a scan date, the referring dentist, the imaging centre and an accession number each tie the scan to the person. The incidental neck finding is health information, which is sensitive information under the Privacy Act, and the patient may not yet have been told. 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 dental radiology report?
Only in part. 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 report would still reach that AI company in full, including the contact between tooth 38 and the nerve canal, the carotid finding and the recommendation to discuss risk with the patient. Our comparisons with ChatGPT Enterprise and Claude weigh the trade-offs.
How can a dental practice use AI for a dental radiology 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 answer is drawn from the report and the patient record held by the practice. The dentist checks every finding against the report and the record and signs it.
Is an imaging report part of the dental record?
Yes. Section 8.3 of the Code of conduct describes a health record as covering clinical history, findings, investigations, information given to patients, medication and other management. A radiology report is an investigation, and it has to be held securely and free from unauthorised access like the rest of the record.
Can AI summarise a radiology report for the patient?
It can produce plain wording, but the practitioner remains responsible for what the patient is told. Ahpra says practitioners must apply human judgment to AI output. A summary that drops an incidental finding, or softens a recommendation for medical review, is the kind of error a person has to catch against the original.
Who has to look after the images and the report?
The practice holds them as part of the patient's record and is responsible for their security under the Privacy Act, including APP 11. In NSW the Health Records and Information Privacy Act 2002 also has principles on security and on disclosure. Copies sent to an overseas AI service are outside that control.

See Dominion answering from your own documents.

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

Book a demo