§ Canonical reference · Source-reviewed monthly
The living reference to Australia's five-source radiology AI governance landscape: RANZCR Chapter 9, Ahpra AI guidance, TGA AI and medical-device software guidance, ACSQHC AI Clinical Use Guide, and DISR Guidance for AI Adoption. Cite this page when a board, insurer, or medico-legal reviewer needs the current state.
Source-reviewed by hand. Not automatically scraped.
Jump to source · freshness at a glance
Cite this page
Reviewed 6 May 2026
Moirai Governance. (2026). State of Radiology AI Governance. Moirai Health Pty Ltd. https://moirai.health/standards
Public intelligence layer
This surface is source-reviewed, not automatically scraped. The paid Tools surface remains the private system of record.
Signal 01
Active pressure
RANZCR, TGA, Ahpra, ACSQHC, and DISR now create overlapping expectations for safe deployment, intended-use review, accountability, monitoring, and evidence.
Public intelligence layer. Moirai maps the landscape; practices still need their own legal and clinical governance review.
Signal 02
Measured, not guessed
AI use is moving from pilots into radiology workflows, but governance maturity is uneven. The TGA AI-enabled ARTG list is now an important public signal, not a complete market map.
Source-reviewed and not automatically scraped. Full customer benchmarks require paid-practice data and permissioned aggregation.
Source confidence
Medium. Public regulatory lists and customer benchmark data still need scale.
Check sourceSignal 03
Private readiness model
Readiness, Assurance, and Navigation scores translate public obligations into concrete evidence gaps, owner actions, review cadence, and board-reportable posture.
The score is a Moirai operating model, not a regulatory certification and not legal advice.
Signal 04
Limited public layer
Public vendor and ARTG references belong here. Private tool records own the customer's full risk posture, incidents, evidence, and monitoring record.
Public preview only. Full vendor assessment, known incidents, and customer evidence remain inside the paid operating file.
Source confidence
Source-backed where public entries exist; otherwise explicitly unknown.
Check sourceSource registry
The register below is generated from the official source registry. It shows what was checked, when it was checked, and the next review date without pretending to run automated legal monitoring.
Last source review
6 May 2026
Official sources
5
High confidence
5
Next review
2026-06-05
Source-reviewed public intelligence; not automatically scraped
Professional standards · Australia and New Zealand
High. Official college standards page.
https://www.ranzcr.com/advocacy-standards/artificial-intelligence/Source status
Professional obligations guidance · Australia
High. Official regulator guidance.
https://www.ahpra.gov.au/Resources/Artificial-Intelligence-in-healthcare.aspxSource status
Medical device regulation guidance · Australia
High. Official regulator guidance, with explicit scope limits.
https://www.tga.gov.au/products/medical-devices/software-and-artificial-intelligence-ai/manufacturing/artificial-intelligence-ai-and-medical-device-software-regulationSource status
Clinical safety guidance · Australia
High. Official safety and quality body guidance.
https://www.safetyandquality.gov.au/clinical-topics/digital-health/artificial-intelligenceSource status
Whole-of-organisation AI governance guidance · Australia
High. Official government guidance.
https://www.industry.gov.au/publications/guidance-for-ai-adoption/ai-register-templateSource status
Recent reviews
Source-reviewed by hand on a fortnightly-to-monthly cadence. Every re-check lands here, dated, with a one-line summary of what actually changed in the public material.
Confirmed DISR's AI register template wording still recommends accountable persons per AI system. No structural change to local Tools record schema required.
ACSQHC AI Clinical Use Guide page updated 20 March 2026. Medical Image Interpretation scenario language tightened around automation bias and post-incident review.
TGA AI-enabled ARTG list expanded with two new entries in radiology category. Re-checked intended-use language; existing buyer guidance still applies.
Ahpra published a clarifying FAQ on practitioner accountability for AI-assisted decisions. Wording on this page already covers it; cross-link added.
Confirmed RANZCR's AI-specific standards still sit inside the broader Standards of Practice. No new chapter; clarified version-change review trigger.
Document 01
Checked 16 days agoReview in 15 daysRoyal Australian and New Zealand College of Radiologists
Current AI standards page, reviewed May 2026
Read sourceOfficial source citation
Source checked 6 May 2026. External reference: https://www.ranzcr.com/advocacy-standards/artificial-intelligence/
Professional standards context for AI use in clinical radiology. RANZCR states that AI-specific standards for clinical radiology sit inside its broader Standards of Practice and should be interpreted with the wider practice standards.
What the source emphasises
How Moirai maps evidence
Document 02
Checked 16 days agoReview in 15 daysAustralian Health Practitioner Regulation Agency and National Boards
22 August 2024, regularly updated guidance
Read sourceOfficial source citation
Source checked 6 May 2026. External reference: https://www.ahpra.gov.au/Resources/Artificial-Intelligence-in-healthcare.aspx
Practitioner-level guidance on professional obligations when using AI in healthcare. It emphasises accountability, human judgement, intended use, limitations, data handling, transparency, consent where relevant, local governance, and indemnity.
What the source emphasises
How Moirai maps evidence
Document 03
Checked 16 days agoReview in 15 daysTherapeutic Goods Administration
AI guidance updated 5 February 2026
Read sourceOfficial source citation
Source checked 6 May 2026. External reference: https://www.tga.gov.au/products/medical-devices/software-and-artificial-intelligence-ai/manufacturing/artificial-intelligence-ai-and-medical-device-software-regulation
Australian medical device regulation is technology-neutral and based on intended purpose. Software and AI can be regulated as medical devices when intended for diagnosis, prevention, monitoring, prediction, prognosis, treatment, investigation of anatomy or physiological processes, or related therapeutic purposes. TGA now also publishes an AI-enabled ARTG list, with stated limitations.
What the source emphasises
How Moirai maps evidence
Document 04
Checked 16 days agoReview in 15 daysAustralian Commission on Safety and Quality in Health Care
AI resource page updated 20 March 2026
Read sourceOfficial source citation
Source checked 6 May 2026. External reference: https://www.safetyandquality.gov.au/clinical-topics/digital-health/artificial-intelligence
Pragmatic AI guidance for health professionals and safety scenarios for clinical AI use, including medical image interpretation. The Commission highlights that AI can introduce new patient risks if not applied thoughtfully and appropriately.
What the source emphasises
How Moirai maps evidence
Document 05
Checked 16 days agoReview in 15 daysDepartment of Industry, Science and Resources
Guidance published October 2025
Read sourceOfficial source citation
Source checked 6 May 2026. External reference: https://www.industry.gov.au/publications/guidance-for-ai-adoption/ai-register-template
Whole-of-organisation AI adoption guidance. DISR recommends maintaining an AI systems register that records AI systems, use cases, accountable people, and appropriate governance level.
What the source emphasises
How Moirai maps evidence
§ Beyond regulatory
Regulatory mapping helps you prepare for review. Security architecture and data sovereignty practices keep you defensible the rest of the time.