Issue 2 · 23 April 2026 · Fortnightly | AI in Healthcare · Australia & Beyond
Summary
Read time: ~5 min
Top Stories
- Australia's Aged Care AI Boom — and the High-Risk Warning That Comes With It
- Australia's First GP Study: AI Scribes Match — and Sometimes Beat — Human Documentation
- AHPRA's 2031 Strategy Puts AI Front and Centre — Including Scanning Clinician Social Media
- JAMA: Ambient AI Scribes Cut EHR Time by 13 Minutes — Across Five Real Health Systems
Signals
- My Health Record FHIR Gateway v1.4 requirements published
- Dept of Health publishes formal AI Transparency Statement
- AHPRA: professional obligations when using AI — accountability, consent, transparency
- AI in Health Regulation Conference (Sydney) concludes with no TGA consensus on AI scribes as devices
Top Stories
Australia's Aged Care AI Boom — and the High-Risk Warning That Comes With It
🇦🇺 Australia | Aged Care · ★ Rapid uptake · expert safety warnings dominate headlines
Australia's aged care sector is rapidly adopting AI — from monitoring sensors and robot companions to predictive care tools and AI-powered rostering systems. ABC News reports the technology is already operating in facilities across the country, raising the quality of routine monitoring and freeing overworked staff from time-consuming observation tasks.
But experts are sounding the alarm. Researchers and clinicians warn that deployment is outpacing governance, with providers adopting tools not validated for frail elderly populations, lacking transparency on decision-making, and operating without clear accountability frameworks. The "safety-by-design" principle — building ethics and oversight in from the start — is largely absent.
"We're seeing uptake happen faster than the evidence, faster than the regulation, and faster than the training." For a sector already under royal commission scrutiny, that's a risk that cannot be ignored.
What it means for clinicians:
- Allied health professionals in aged care need to ask hard questions of vendors: Is this tool validated for aged populations? What happens when it gets something wrong?
- Documentation of AI use and clinical oversight is increasingly a professional and legal expectation — not optional.
- AHPRA's upcoming governance frameworks are likely to extend into aged care, making now the time to audit which AI tools you're using and why.
Australia's First GP Study: AI Scribes Match — and Sometimes Beat — Human Documentation
🇦🇺 Australia | Clinical Research · ★ First published AU GP AI scribe study · RACGP AJGP April 2026
The Australian Journal of General Practice has published the first study comparing AI scribes to human documentation in a simulated Australian general practice context. Researchers tested four commercially available AI scribes (out-of-the-box) against a human scribe across RACGP Fellowship exam cases. Three experienced GPs rated all outputs blind.
Results: AI scribes demonstrated comparable or superior performance across accuracy, thoroughness, and succinctness. The overall difference was not statistically significant (P=0.071) — but AI scored significantly better on thoroughness (P<0.001) and succinctness (P<0.001). More confronting: both AI and human documentation contained "hallucinations," challenging the assumption that human notes represent the gold standard.
The study was simulated, not real-world, and each scribe showed distinct strengths and weaknesses — making thoughtful product selection critical. But the core finding stands: commercially available AI scribes already produce documentation that experienced GPs cannot reliably distinguish from their own work.
What it means for clinicians:
- The "AI makes things up" objection is weakened — human notes contain hallucinations too. The question is which errors matter more in context.
- Allied health professionals should test multiple AI scribe tools before committing — performance varies significantly between products.
- Australian-specific validation matters: a tool trained on US or UK data may perform differently in local clinical contexts.
AHPRA's 2031 Strategy Puts AI Front and Centre — Including Scanning Clinician Social Media
🇦🇺 Australia | Regulation · ★ National Scheme Strategy 2031 published · March 2026
AHPRA has published its National Scheme Strategy 2031, with AI featuring prominently. The regulator identifies AI, telehealth, and evolving technologies as one of three forces reshaping health regulation over the next five years — alongside workforce strain from an ageing population and the rise of health self-sourcing.
Most significantly: AHPRA has confirmed it is already testing AI to scan publicly available online content — including clinician websites and social media advertising — to flag regulatory concerns. A single practitioner identifier is also planned (from study through to retirement), and the regulator anticipates increased notifications as AI increases the "contestability" of practitioner decisions.
AHPRA stressed that AI "doesn't make decisions" and all regulatory outcomes remain human-made. But the direction of travel is clear: the regulator is building AI into its own operations while expecting registered practitioners to account for AI in theirs.
What it means for clinicians:
- Your online presence — including social media — may already be monitored by AHPRA AI systems. Review your public-facing content now.
- The expected rise in notifications means clinical documentation quality matters more than ever. Thorough, accurate records are your primary defence.
- AHPRA's governance expectations for AI use will only become more codified under the 2031 strategy — start building compliant workflows now.
JAMA: Ambient AI Scribes Cut EHR Time by 13 Minutes — Across Five Real Health Systems
🌏 Global | Clinical Evidence · ★ JAMA multi-site study · 5 academic medical centres
A peer-reviewed JAMA study has quantified ambient AI scribe impact across five US academic medical centres — the most rigorous multi-site evidence to date. Using Ambience, Nuance Dragon DAX Copilot, and Abridge (all with Epic), researchers found AI scribes decreased total EHR time by **13.4 minutes per clinician and documentation time by 16.0 minutes** per session.
Productivity gains were material: clinicians using AI scribes saw 0.49 more patient visits per week — a meaningful capacity increase in stretched health systems. The AHA profiled six health systems now operating at scale with ambient scribes across oncology, emergency medicine, and outpatient settings.
Where previous studies relied on single-site pilots, this JAMA publication provides multi-centre, real-world data at meaningful scale. For Australian health systems still evaluating the technology, this study significantly raises the evidence bar.
What it means for clinicians:
- **13 minutes saved per session compounds fast. For a full-time clinician seeing 20 patients a day, that's over 4 hours** recovered per week.
- The additional 0.49 visits per week per clinician suggests ambient scribes increase practice capacity — not just reduce burnout.
- EHR integration was central to the efficiency gains. Australian clinicians should prioritise tools that integrate with their existing PMS.
Signals
AU | Digital Health — The Australian Digital Health Agency published v1.4 of the My Health Record FHIR Gateway Operations Requirements (DH-4308:2026), providing clearer technical standards for apps integrating with Australia's national health record. Essential reading for any health tech vendor building on the My Health Record ecosystem. Read →
AU | Policy — The Department of Health and Aged Care published a formal AI Transparency Statement detailing how AI is used within the department — including a use-case register, staff training protocols, and a commitment to human oversight of all AI-assisted decisions. Sets the governance tone for the broader health sector. Read →
AU | Clinicians — AHPRA's professional obligations guidance for AI in practice remains essential: accountability stays with the clinician regardless of TGA approval, informed consent must be documented, and practitioners must understand how their AI tools were trained and on what populations — before using them. Read →
AU | Events — The AI in Health Regulation, Policy and Standards Conference (Sydney, 31 March–1 April) concluded without regulatory consensus on whether AI scribes qualify as TGA-regulated medical devices. Compliance clarity for clinicians is still months away — watch this space. Read →