Further information about AI

Artificial intelligence or AI can be defined as:
‘computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision making, and translation between languages¹’

There are different types of AI including:

• Machine Learning - is a subset of AI that allows computers to autonomously learn and improve without being explicitly programmed. Machine Learning algorithms are trained on data to make predictions or decisions. For example, diagnostic tools that can enhance image scanning, support decision making and predict the risk of disease
• Generative AI is a type of Machine Learning AI – creates new content such as text, images, voice, video and code by learning from data patterns. Examples are ChatGPT and Google Bard. AI scribing tools are also being developed specific for healthcare use, that enable health practitioners to automate parts of making clinical records by converting a conversation with a patient into a clinical note, summary, or letter that can be incorporated into the patient’s health record.
• Natural Language Processing (NLP) - is a field of artificial intelligence that deals with the ability of computer systems to understand and generate human language. NLP algorithms are used to analyse text, comprehend, converse with users and perform tasks like language translation, sentiment analysis, and question answering, and
• Computer Vision (CV) - empowers computers to 'see' and comprehend the visual world, analysing images and videos like humans. CV algorithms analyse images and videos for tasks like object detection, face recognition, and self-driving cars¹.

¹ www.digital.nsw.gov.au/policy/artificial-intelligence/a-common-understanding-simplified-ai-definitions-from-leading; A common understanding: simplified AI definitions from leading standards; Digital.NSW; NSW Government; Accessed 8 March 2024

The TGA regulates AI-enabled products when they have a therapeutic use and meet the definition of a medical device unless otherwise exempt or excluded.

More information about when AI is considered a medical device and regulation of generative AI is available on the TGA’s website.

Find out more

• When AI is considered a medical device
• Artificial intelligence chat, text, and language
• Generative AI regulation
  Large language models and chatbot regulation.
• Off-label use
• Excluded software
  There are 15 conditional software 'exclusions' within the Therapeutic Goods (Excluded Goods) Determination 2018, that help clarify the boundary of products that are used within healthcare but are not considered to meet the definition of a medical device.
• Intended purpose of software
  Manufacturers are required to provide evidence to support the intended purpose of a medical device (including software and AI).
• Exemption for Certain Clinical Decision Support Software
  Software (including Clinical decision support system (CDSS) software) that makes suggestions for diagnosis or treatment is subject to regulation by the TGA and must undergo pre-market approval and be included in the Australian Register of Therapeutic Goods (ARTG).
  Certain low risk CDSS software may be exempt from inclusion in the ARTG under specific conditions that must be met.
• Software-based medical devices FAQs

There are other legislative requirements that practitioners need to consider when using AI including (but not limited to) requirements from authorities that regulate the capture and storage of health information, including heath records, and privacy legislation and/or any other relevant privacy requirements.AI products that do not fall within the TGA regulatory framework must still comply with consumer and advertising law. Further information is available at:

• Advertising software based medical devices to healthcare professionals (PDF)
• Advertising software based medical devices to consumers (PDF)
• Cyber security guidance for industry
• Cyber security information for users
• Cyber security information for consumers

There is a range of information and resources about AI available. The list below is not intended to be exhaustive.

Some state and territory health and other government departments have information and resources about AI.

• Australian Commission on Safety and Quality in healthcare
• Office of the Victorian Information Commissioner - information about Understanding Privacy Obligations when using Artificial Intelligence
• NSW health information about Clinical applications of artificial intelligence
• A Roadmap for AI in Healthcare for Australia | Ai Health Alliance

Many professional associations and professional indemnity providers have published information and resources about AI. Some recent publications include:

• Using Artificial Intelligence (AI) tools for record management in doctor consultations – MDA National
• Australian Dental Association - Policy Statement on Artificial Intelligence in Dentistry
• Artificial Intelligence in Healthcare | Australian Medical Association (ama.com.au)
• RACGP guidance on AI scribes
• ASMIRT releases AI Position Statement – ASMIRT
• The Cyber Impacts of AI in Nursing and Midwifery (anmf.org.au)
• APA | Innovation at hand: AI's impact on physiotherapy (australian.physio)
• Australian College of Nursing – Position Statement Artificial Intelligence
• ACD Position statement – Use of Artificial Intelligence in Dermatology in Australia
• RANZCR Ethical Principles for AI in Medicine
• RANZCR Position statement – Regulation of Artificial Intelligence in Medicine