Ombudsman finds uses of automation increasing across NSW govt


A new report from the NSW Ombudsman tabled in Parliament earlier in the month found the NSW public sector has exponentially increased its use of automated decision-making (ADM) systems since a previous report released in 2021.

The new report indicated that the lack of visibility and obligation to release this information to the public among government departments and agencies related to their uses of ADM systems, including artificial intelligence (AI) tools, was a concern.

It also provided NSW’s first map of how its public sector is using or planning to use ADM systems to perform certain functions and navigate the associated risks, with 275 systems currently identified. It also presented a standard definition of an ADM system: “a fully or partially automated technical system, used by a NSW government organisation (state government department or agency, or local council), in administrative decision-making, and that affects people”.

The research, conducted by the ARC Centre of Excellence on Automated Decision-Making and Society (ADM+S), brings together in one source information gathered from participating agencies, local councils and publicly-available information about the use of ADM systems in the NSW state and local government sectors.

“We believe that members of the public whose rights and interests have been materially affected by a decision made with the use of ADM, are entitled to be informed of the role ADM played in that decision,” NSW Ombudsman, Paul Miller, said.

“Visibility is necessary for people to properly consider and exercise any decision review rights as well as for proper oversight. It is also key to supporting an informed debate about what assurance and regulatory frameworks may be appropriate for ADM use now and into the future.”

The research showed that one-third of the reported systems were “in development, being piloted or planned within the next three years”, mostly by NSW local councils. While government agencies were interested in implementing ADM systems using forms of AI, including predictive analytics, natural language processing and generative AI, simpler technologies and automation formats were currently more widespread and “heavily relied upon within government”.

Across a total of 11 departments and 195 affiliated agencies, there were 136 ADM systems reported being used, with the Transport portfolio and Communities & Justice portfolio recording the highest number of systems used.

“Although this mapping project has produced a point-in-time overview of ADM usage, we echo the research team’s view that voluntary public reporting will not be sufficient to ensure comprehensive and continuing visibility of ADM use across the public sector going forward,” Miller said.

“We hope that all departments, agencies and local councils that have contributed to this research will find the report analysis and insights of value, and useful as they continue to consider and pursue their own current and future ADM projects.”

“This project has been a unique and really exciting opportunity to provide some transparency around how automated decision-making systems are being used in government in NSW,” Professor Kimberlee Weatherall, Professor of Law, The University of Sydney Law School and Chief Investigator, ARC Centre of Excellence on Automated Decision-Making and Society, said.

“We’ve found that automation is widespread and increasing, across both state government and local councils. Our hope is that we’ve provided useful insights into where automation is at in NSW, as well as a basis for a better informed, ongoing conversation about automation and artificial intelligence in government.”