How do you design a trustworthy system?
We help our clients build fairness, ethics, accountability and transparency (‘FEAT’) into AI projects, algorithmic systems and automated decision-making.
Our Algorithmic Impact Assessments use the Four D’s Framework to assess privacy risk, across design, data, development, and deployment.
For an assessment of an algorithmic system to be robust, it should encompass:
- Legal compliance – ensure the algorithmic system is lawful, with particular focus on privacy, anti-discrimination, and consumer protection laws
- Social impacts – consider the social, political, and economic context for a deeper appreciation of potential privacy-related harms, and
- Technical considerations – integrate testing for accuracy, performance, fairness and bias.
For more about the Four D’s Framework, see Algorithms, AI, and Automated Decisions – A guide for privacy professionals.
Clients for whom we have provided advice on AI projects include:
- Ambiata – part of the Insurance Australia Group, Ambiata is a leader in Australia’s AI industry and a pioneer in the use of machine learning at-scale within the Australian market. We conducted a PIA for Ambiata on its development of a new platform for secure data-sharing and analytics, viewed through the lens of different potential use cases
- Victorian Department of Justice and Community Safety – advice on the Project Kingfisher major procurement project including development of a bespoke set of AI and privacy-related guardrails for infringement management and enforcement, as relevant to contact centre and BPO functions, to form the basis of the Department’s tender requirements
- Cancer Institute NSW – PIA on the BreastScreen Machine Reading Evaluation project to trial the use of AI technology to ‘read’ mammography images
- NSW Ministry of Health – PIA on the adoption of AI tools designed to improve the accuracy and efficiency of clinical documentation and clinical coding processes; and
- EWOSA – a PIA for the Energy and Water Industry Ombudsman SA on a customised, internal AI tool to assist EWOSA’s dispute resolution officers to investigate and manage customer complaints.
