Leichhardt General Practice Policy for Use of Artificial Intelligence (AI)
Correct as of 16 / 07 / 2026
Introduction
This policy outlines the safe, ethical, and compliant use of Artificial Intelligence (AI) tools within Leichhardt General Practice, ensuring alignment with the Royal Australian College of General Practitioners Standards for General Practices (5th Edition).
The policy ensures:
Patient safety and quality care are maintained
Confidentiality and privacy obligations are upheld
AI is used as a support tool, not a decision-maker
Scope
This policy applies to:
All GPs, registrars, nurses, allied health professionals
Administrative and reception staff
Contractors and students
It covers all AI tools, including:
Clinical decision support systems
Documentation/scribing tools
Generative AI (e.g., ChatGPT-style tools)
Automation software used in administrative workflows
Definitions
Artificial Intelligence (AI):
Software capable of performing tasks that typically require human intelligence, including data analysis, text generation, and clinical support.
Generative AI:
AI systems that create content such as clinical notes, summaries, or patient communications.
Clinical Decision Support (CDS):
AI tools that assist but do not replace clinician judgement.
Policy Statement
Leichhardt General Practice supports the responsible and controlled use of AI to enhance efficiency and patient care, provided that:
AI outputs are always reviewed by a qualified clinician
No AI tool replaces clinical judgement or accountability
Patient data is protected in accordance with Australian privacy law
AI use aligns with RACGP Standards, including:
C2 – Clinical Governance
C6 – Information Security
QI1 – Quality Improvement
Principles for AI Use
Patient Safety First
AI must not be used to make autonomous clinical decisions
All outputs must be clinically verified before use
Human Oversight
The treating GP remains fully responsible for:
Diagnosis
Treatment decisions
Documentation accuracy
Transparency
Patients should be informed if AI is used in:
Documentation (e.g., scribing tools)
Communication or summaries
Privacy and Confidentiality
No identifiable patient data is to be entered into:
Public AI tools
Non-compliant platforms
Unless:
The platform is approved, secure, and compliant with:
Australian Privacy Principles (APPs)
RACGP data security requirements
Acceptable Uses of AI
AI may be used for:
Clinical Support (with oversight)
Drafting clinical notes
Summarising patient histories
Generating patient education materials
Administrative Efficiency
Appointment summaries
Internal documentation
Policy drafting
Quality Improvement
Data analysis (de-identified)
Audit support
Prohibited Uses
AI must NOT be used for:
Autonomous diagnosis or prescribing
Uploading identifiable patient data into unsecured systems
Replacing clinical consultations
Generating medico-legal documents without review
Any use that compromises patient confidentiality
Data Security and Privacy
In alignment with RACGP C6 – Information Security:
Only approved AI tools may be used
All tools must meet:
Secure data storage requirements
No unauthorised offshore data transfer
Staff must not:
Copy/paste patient-identifiable data into public AI tools
Breaches must be reported under the practice’s:
Data breach policy
Incident management system
Clinical Governance and Risk Management
Aligned with RACGP C2 – Clinical Governance:
AI use is subject to:
Ongoing monitoring
Risk assessment
The practice will:
Maintain a register of approved AI tools
Review AI outputs periodically
Audit usage where appropriate
Staff Training and Responsibilities
All staff must:
Complete training on:
Safe AI use
Privacy obligations
Understand limitations of AI (bias, hallucinations, errors)
Use AI tools only within their scope of practice
Patient Consent
Where AI is used in patient care:
Verbal consent should be obtained if:
AI is directly involved in documentation or communication
Patients may opt out
Documentation Standards
AI-generated content must be:
Reviewed
Edited as required
Clearly attributable to the clinician
Medical records must meet RACGP requirements for:
Accuracy
Completeness
Timeliness
Incident Management
Any AI-related issue must be reported, including:
Clinical errors linked to AI
Data breaches
Misuse of AI tools
These will be managed under the practice’s:
Incident reporting system
Continuous quality improvement framework
Continuous Improvement (QI1)
The practice will:
Regularly review AI use
Incorporate feedback from staff and patients
Update policy in line with:
Regulatory changes
Emerging risks
Review Schedule
This policy will be reviewed annually, earlier if:
New AI technologies are introduced
Regulatory requirements change