Skip to content

The role of artificial intelligence in regulatory decision-making

9 October 2020

News and media


Artificial intelligence vector image


This guest editorial was written for the ANZSOG/National Regulators Community of practice monthly newsletter, highlighting new additions to the Regulation Policy and Practice collection on APO. The RP&P collection brings together a range of practical resources from national, local and state/territory governments, regulatory agencies and external institutions conducting monitoring, inquiries and reviews. You can receive this newsletter by joining the ANZSOG/National Regulators Community of Practice (membership is free) or subscribe to the newsletter directly.

Many of us already engage with artificial intelligence (AI) and machine learning technologies every day. We use virtual voice assistants, check travel times with online maps and find, when we search the web, that many additional options are suggested to us based on our previous activities and those of others like us. Sooner than we think, we may have self-driving cars.

While there have been numerous reports on the need to regulate AI in the interests of the community, these and related new technologies also offer an opportunity for regulators to transform their internal processes. The 2019 ANZSOG/NRCoP forum The Intelligent Regulator featured renowned speaker Bill Eggers from Deloitte Global Insights talking about several of the new tools available to regulators.

AI, in particular, can help us meet our aspirations for contemporary best practice in continuous review and improvement of the stock of regulation. It can rapidly sift through vast quantities of legislation, map out complex legal landscapes and yield insights to help regulators assess whether rules remain fit-for-purpose.

Used appropriately, AI and machine learning technologies can help us make better informed decisions in less time and drive significant efficiencies. Yet, to date, their application in the context of government regulation has been limited, notwithstanding the recent work of CSIRO in setting out an AI roadmap for Australia.

NSW Treasury recently published a research report, Regulating for NSW’s Future, that draws on data and insights using AI and text-analysis software to assess the NSW regulatory landscape. The software was used to identify regulatory improvement opportunities that could facilitate economic recovery post-COVID-19.

The AI tool analysed the volume, interdependencies and frequency of reviews and amendments to NSW Acts and Regulations. It also pinpointed sections of legislation that contained outdated, onerous or prescriptive language. Using it, NSW Treasury identified opportunities to:

Modernise outdated compliance requirements: the AI software highlighted 453 sections of NSW regulation that require the use of facsimile, telegram and cassette.
Replace prescriptive rules with outcomes-focused regulations: prescriptive language (such as ‘must’, ‘cannot’ and ‘ought’) was found in 37 per cent of sections.
Reduce regulatory overlap and duplication: cafes and restaurants, for example, must comply with 103 sections of regulation enforced by 14 regulators.

Lessons learned

NSW Treasury’s report highlights how useful AI technology can be in accessing regulatory data and how it provides insights without labour-intensive processes. It also emphasises that government agencies have a critical role to play in evaluating AI findings to inform regulatory decision-making. AI is a tool rather than a solution.

While AI can identify regulations that have remained unchanged for some time, age alone does not indicate a need for change. For example, criminal laws, qualification requirements for doctors and road safety rules that were introduced a long time ago are still relevant and necessary today.

Similarly, AI can show the volume of regulations applying to each industry sector. This can provide useful information on sectors that would benefit most from red tape reduction initiatives such as the removal of overlapping requirement. Regulators must, however, assess the costs and benefits in removing specific compliance requirements.

AI is no longer tomorrow’s technology; it is here today, and presents a valuable opportunity to modernise regulatory decision-making.

Scott Wheeler is currently Director of the Economic Strategy Division at NSW Treasury and a member of the NSW chapter committee of the ANZSOG/NRCoP. He has 30 years of experience in public policy, economic analysis and management, working in various positions in NSW Treasury over 17 years as well as in London for the Australian High Commission and in Canberra in both the Treasury and in the Offices of the Treasurer and Prime Minister.