The short version: In August 2024, AHPRA published guidance for all registered health practitioners on using AI in clinical and administrative practice. The guidance applies to every health profession — dental, medical, allied health, nursing, pharmacy, psychology and more — and to every AI tool deployed in practice, including AI receptionists, automated booking systems, and medical scribing tools. The core message is that practitioners remain accountable for AI outputs, must understand the tools they use, and must inform patients about AI involvement. This article unpacks what that means in practical terms for your clinic's front desk.
On 22 August 2024, AHPRA and the 15 National Boards jointly published "Meeting your professional obligations when using Artificial Intelligence in healthcare." It applies to every registered health practitioner in Australia and explains how existing Code of Conduct responsibilities translate to AI use.
The guidance is short — under three pages of substantive content — but the implications are wide. It is also designed to be updated regularly, which means clinics deploying AI tools need a process for monitoring AHPRA guidance changes, not just a one-off compliance check.
Importantly, AHPRA explicitly acknowledges AI's potential to reduce administrative burden and burnout, and signals openness to its adoption. The guidance is not anti-AI. It is pro-accountability.
AHPRA's guidance is built around five principles. Each one applies to AI receptionist deployment as much as to clinical AI use.
The registered practitioner remains fully responsible for any AI deployed in the course of their practice. This applies regardless of whether the AI is making a clinical decision or an administrative one. If your AI receptionist routes a patient to the wrong practitioner, books them at the wrong time, or fails to escalate an urgent clinical question, the responsibility sits with you — not the AI vendor.
What this means at the front desk: One named practitioner in the practice must be the accountable owner of the AI receptionist system. They sign off on the call scripts, the routing logic, the escalation rules, and the data flows. They review the system regularly and document changes.
Practitioners must understand the capabilities and limitations of the AI tools they deploy. AHPRA specifically warns against deferring to AI outputs without applying professional judgment. The practitioner cannot say "the AI handled it" as a defence — they must be able to explain what the AI does, what it does not do, and what its known failure modes are.
What this means at the front desk: Get a written summary from your AI receptionist vendor covering: what the AI is configured to do, what it is configured not to do, what happens when it encounters something out of scope, and what the escalation pathway is. Keep this on file. Train your practitioners and team on how it works before go-live.
Patients must be informed about AI involvement in their care. AHPRA notes that the level of disclosure should match the extent of the AI's involvement. For AI receptionists, this means patients should know they are speaking with an AI when they call the practice — not be misled into thinking they are speaking with a human staff member.
What this means at the front desk: The AI receptionist identifies itself at the start of every call. The clinic's privacy policy discloses the use of AI in patient contact, intake, and booking. From 10 December 2026, this disclosure becomes mandatory under Privacy Act 2026's APP 1.7 as well. Build the compliant disclosure once and it covers both AHPRA and Privacy Act obligations.
Where AI involves the use of patient personal information, patients should consent to that use. AHPRA's wording is that practitioners "need to involve patients in the decision to use AI tools that require input of their personal patient data and if a patient's data is required for care."
What this means at the front desk: Your patient intake forms and consent forms should reference AI use. For new patients, this is typically a single line in your standard consent: "We use an AI receptionist to handle initial calls and book appointments. Your contact details and reason for visit are processed by the system and stored in our practice management software. You can request to speak directly with a human staff member at any time." For existing patients, an updated privacy notice at their next visit usually suffices.
Practitioners must address the broader legal and ethical context of AI use — privacy, data security, advertising compliance, professional indemnity, and TGA requirements where the AI tool is regulated as a medical device.
What this means at the front desk: Most AI receptionists do not qualify as medical devices under TGA rules because they perform administrative rather than clinical functions. But you still need to confirm: (a) where patient data is stored (Australia, ideally); (b) that your professional indemnity insurer is aware of the AI tool; (c) that any advertising claims about your AI use are accurate and not misleading; and (d) that your data security obligations under APP 11 are met.
Before go-live, every AHPRA-compliant AI receptionist deployment should have four documents in place:
Once live, AHPRA compliance is mostly about discipline:
"AHPRA only cares about clinical AI, not admin AI." The guidance applies to all AI used in the course of practice. While clinical AI carries higher risk, administrative AI is not exempt. The accountability, understanding, transparency, and consent obligations apply regardless.
"The AI vendor is liable if something goes wrong." Not under AHPRA's framework. The practitioner remains accountable. Your contract with the AI vendor can allocate commercial risk, but it cannot transfer your professional accountability.
"AI receptionists are too risky for clinics." The opposite is true when deployed correctly. Manual reception introduces its own risks — missed urgent calls, inconsistent intake, untrained casual staff, no audit trail. An AHPRA-aligned AI receptionist with proper escalation logic, consistent scripting, and complete call records often improves the compliance posture compared to manual handling.
"We can wait until enforcement starts." The Privacy Act 2026 deadline (10 December 2026) creates a hard milestone for transparency obligations that overlap heavily with AHPRA's transparency principle. Most clinics will need to update their privacy policy and patient consent language regardless. Doing it once, properly, covers both regimes.
AHPRA's AI guidance does not exist in isolation. For Australian clinics deploying AI at the front desk, three regulatory regimes overlap:
A well-designed deployment satisfies all three through a single set of documents and operational practices. We've put the full compliance picture into a downloadable checklist below.
One-page practical checklist covering AHPRA AI Guidance, Privacy Act 2026 (APP 1.7), ACMA telemarketing standards, and Fair Work Right to Disconnect. Same document we use internally for every build at RJ Does AI.
Download the checklist