AI Safe
Artificial Intelligence (AI) is an emerging technology with a wide range of applications in health, data analysis and service delivery. A key issue for healthcare providers like the Foundation is to understand when and how it is appropriate to use AI. At present, AI can replicate human decision-making processes at a large scale but lacks the ethical or moral framework, or contextual awareness, necessary for good decision making. As AI is designed and programmed by humans, there is a significant risk that it amplifies inherent biases present in all human beings. This page sets out the Foundation’s policy on how it will use AI, and the ethical framework it places around the application of AI in the provision of health care and services to the community..
AI Safe Label
When you see this label on a program, video or other service it means that AI has been used in production. Under our policy framework, the Foundation will inform you how AI was used so you are fully informed.
AI Ethical Principles
AI must clearly be the best solution that maximises the benefit for individuals and/or the community
Other solutions have been considered and ruled-out because they will not realise the same or similar benefits of an AI solution.
Use of AI will include safeguards to mitigate the substantial risk of bias or data quality.
-
AI decision maker is not generalised from a subset of individuals when this does not account for the unique variables present in a whole community.
-
Datasets are genuinely representative for the problem to be solved.
-
Regular human monitoring and quality assurance of data models and outputs is undertaken.
-
Privacy and confidentiality will be maintained, and community members advised when and if AI is being used.
-
At all times AI initiatives are overseen by individuals with the relevant expertise in the technology and its benefits and risks.
AI Ethics Policy
​
The Foundation’s Ethical Use of AI Policy will be published here shortly.
AI Uses
A list of the limited cases where the Foundation uses AI will be available here shortly with detailed information.
