Australian GPs Should Move Cautiously With AI in Mental Health Care
Australian general practitioners should introduce artificial intelligence into mental healthcare with explicit patient consent, clinical safeguards and ongoing oversight, according to a new article in the Australian Journal of General Practice. The paper examines how AI could expand access to mental health services while identifying the risks that require careful management.
The Access Problem
Mental healthcare remains a major access challenge in Australia. An estimated 42.9% of Australians aged 16-85 experience a mental disorder in their lifetime, yet fewer than half of those needing care receive treatment.
AI tools now available to GPs include clinical documentation systems, digital therapy platforms, chatbots, just-in-time support tools, predictive analytics and agentic AI models. These could help make care more responsive and personalised-supporting patients between consultations and reducing administrative work for clinicians.
Support, Not Replacement
"AI should support, not replace, clinical judgement," the authors said. Given the sensitive and personal nature of mental healthcare, careful integration is essential.
GPs must ensure patients know when AI is being used, what role the tool plays, what information it collects and what privacy risks may exist-particularly when systems access clinical data or interact directly with patients.
Where to Start
The authors recommended beginning with lower-risk use cases: stable patients with mild-to-moderate symptoms rather than patients in crisis or those with complex psychiatric conditions.
AI tools should function as part of a broader care plan. GPs must continue reviewing patient progress and checking AI-generated material-including consultation summaries-for omissions, errors or misleading interpretations. Clinicians must retain the ability to challenge or override AI outputs when they don't fit the patient's circumstances, preferences or clinical context.
Accountability and Regulation
Clinical responsibility remains with the GP when AI outputs inform care decisions, regardless of whether a tool is formally regulated as a medical device. Different tools may be subject to different regulatory levels depending on their function.
Practical barriers also affect adoption. GPs need AI tools that integrate into existing electronic medical record systems and everyday workflows, along with clear funding arrangements-whether through Medicare Benefits Schedule items, private payment or other mechanisms.
What's Missing
More evidence is needed before wider implementation in Australian general practice. Clinical trials and real-world studies comparing AI-supported GP mental healthcare with standard care would help establish safety and effectiveness.
GPs also need training on data ethics, bias, medico-legal responsibility and appropriate use of AI for Healthcare in mental health settings.
Regional Context
The Australian government announced A$135.2 million in investment in digital mental health projects last year. In New Zealand, a national pilot testing Generative AI and LLM systems as scribes in public emergency departments recently expanded its use in emergency mental health services.
Your membership also unlocks:
