Safe and Responsible Artificial Intelligence in Health Care – Legislation and Regulation Review Final Report
Safe and Responsible Artificial Intelligence in Health Care Legislation and Regulation Review Final Report March 2025
This final report reviews Australian health portfolio legislation to identify gaps and vulnerabilities in the regulation of artificial intelligence in healthcare settings. Informed by a public consultation, literature review, and stakeholder engagement, the report outlines findings across seven key areas including regulation, national AI health policy leadership, implementation resources, trusted information, benefits realisation, data and consent strengthening, and incentives. It concludes that while existing health legislation can largely accommodate AI with minor technical amendments, economy-wide guardrails and non-regulatory initiatives are needed to manage risks.
Summary and content fields are drawn from the document text and are pending editorial review.
Who issued it
Who it governs
Policymakers, regulators, healthcare providers, and researchers interested in the regulation and safe adoption of artificial intelligence in Australian health care.
What the document contains
The document covers a review of Commonwealth health portfolio legislation regarding artificial intelligence, findings from public consultations and stakeholder engagement, key themes such as bias, transparency, consent, and data risks, and potential regulatory and non-regulatory initiatives.
The document does not provide binding legal rules or final amendments, as it is a policy review report detailing findings and feedback for future government consideration.
- Current Australian health portfolio legislation is largely able to accommodate artificial intelligence but may require minor technical and definitional amendments for clarity.
- Stakeholders widely support national and centralised policy leadership to steward the equitable and safe implementation of AI in health care.
- Robust data governance and consent strengthening are critical to managing unique risks associated with the high sensitivity of personal health data.
- A trusted, centralised source of high-quality information is needed to support consumers and clinicians in making informed decisions about AI products.
- Maintaining a meaningful human in the loop is viewed as critically important for patient safety and clinical decision-making.
· Overview · Findings · Key Themes · Background · Objectives · Scope · Legislation and regulation · Public consultation · Conclusion · Appendix A – Legislative framework overview · Appendix B - Consultation · References
How it scores
Each score is backed by a quote verified against the document text. Editorial review has not yet confirmed them.
Accountability
4/5Clear responsibility for AI outcomes
Regulation should be clear about when and how data can be accessed and used by AI and who is accountable for the responsible use of patient data.
Finding 6: Data and consent strengthening
screening score · 2026-10-24
Fairness
3/5Non-discrimination and equitable treatment
Most respondents were concerned about the potential for bias, especially where Australian populations are not well represented during the development of AI products.
Bias
screening score · 2026-10-24
Human Oversight
4/5Humans remain in control of consequential decisions
Almost all respondents identified having a human in the loop as being critically important for patient safety.
Human in the loop
screening score · 2026-10-24
Privacy
4/5Protection of personal data and private life
The Privacy Act 1988 establishes the Australian Privacy Principles (APPs), which set requirements for handling personal and sensitive health information.
Whole of economy legislative context
screening score · 2026-10-24
Transparency
4/5Openness about how AI systems work and are used
Stakeholders also raised the need for transparency about when AI is being used in the delivery of care and how the AI product works (including how it derives outputs).
Consent and transparency
screening score · 2026-10-24
Enforcement
2/5Consequences and compliance mechanisms with real force
Existing frameworks do provide avenues of redress for aggrieved individuals in situations where AI is used in health care. However, difficulties in detection and enforcement support enhancing current protections
Legislative review findings
screening score · 2026-10-24
Impact Assessment
4/5Required assessment of an AI system's risks/impacts before or during use
Where gaps do exist (notably health products which fall outside the scope of TGA regulations) economy-wide AI guardrails, incorporating pre-market and post-market requirements may provide clarity
Finding 1: Regulation
screening score · 2026-10-24