Anyone who has studied first year economics will recall the basics of the production theory: when labour input rises, total output should follow. While this sounds simple in theory, Australia’s economic reality has thrown a real curveball into the equation.
[Grace Zhou is a fourth year Law and Commerce student who has a strong passion for exploring the intersection between macroeconomic policy and the law. She is fascinated by the role of regulatory frameworks such as antitrust law on shaping market efficiency, corporate governance and macroeconomic stability. Combining her analytical skills and an understanding of macroeconomic theory, Grace aims to provide digestible insights to students who wish to learn more about current economic trends, and how regulatory policies shape Australia’s economic resilience and drive sustainable growth.]
Productivity Talk at the Economic Reform Roundtable
Anyone who has studied first year economics will recall the basics of the production theory: when labour input rises, total output should follow. While this sounds simple in theory, Australia’s economic reality has thrown a real curveball into the equation. Despite the fact that we are working longer hours, aggregate output has barely seemed to budge. In the year to March 2025, labour productivity fell by 1%, while multifactor productivity crawled up just slightly by 0.1% (Productivity Commission, 2025). This is well below the 20-year average of 0.3% and almost negligible in comparison to the 1.6% annual growth rate we experienced during the 1990s productivity boom (Productivity Commission, 2025). At the Economics Reform Roundtable in August 2025, Australian business leaders and experts wrestled with the concern of the productivity slowdown. Atlassian co-founder Scott Farquhar argued that AI will be a once-in-a-generation productivity lever and urged Australia to seize opportunities in AI-driven innovation and infrastructure across industries (Forbes, 2025). The Australian government seems to agree with Farquhar’s point. The government has identified AI as a critical technology which is capable of driving productivity growth across the economy. According to the Department of Industry, Science and Resources, automation could add up to $600 billion a year to the nation’s GDP by 2030 (Australian Trade and Investment Commission, 2025). Within the infrastructure sector alone, plans are already underway to expand Australia’s digital footprint via an investment of $20 billion between 2025 and 2029 (Australian Trade and Investment Commission, 2025). This article explores how AI is being tested and applied to the infrastructure and transportation sectors, namely in improving road maintenance and optimising traffic control.
Automation in Road Maintenance
Australia’s road networks are under increasing strain. Growing populations, heavier traffic flows and limited council resources are stretching infrastructure to its limits. Between 1998 and 2018, total vehicle kilometres travelled rose by an average of 1.95% per year (Infrastructure Australia, 2025). However, funding has failed to keep pace and alleviate the mounting pressure this has put on existing roads. The Grattan Institute estimates a $1 billion annual shortfall in road maintenance funding (NEC, 2025). This is a gap reflecting not just a financial strain but widening cracks in the safety, efficiency and long-term sustainability of Australian roads. The 2019 Australian Infrastructure Audit confirmed the root causes of this gap: underspending on maintenance, short and inconsistent funding cycles, poor data collection and limited performance reporting (Infrastructure Australia, 2025). Additionally, councils have always relied on manual inspections, community reports and routine maintenance schedules to identify defects across the road network. However, this reactive model is inherently limited as deterioration often goes undetected until it poses a hazard, at which point emergency repairs may become unavoidable. The result is skyrocketing costs and more widespread disruptions to both road safety and road network efficiency.
In light of Australia’s mounting road maintenance challenges, automation may be able to present a smarter way forward. Recent research shows that AI-powered monitoring systems, such as machine learning models like Multi-Layer Perceptron, Support Vector Machine and Random Forest, can analyse road-surface data and detect pavement deterioration with up to 98.98% accuracy (Bibri et al., 2023). If public transport such as buses and trams, along with commercial freight trucks, are fitted with cameras and vibration sensors, road authorities can continuously collect real-time data on cracks, potholes and surface wear. AI algorithms can then process this information and identify subtle deteriorations long before it becomes visible to human inspectors. When integrated with weather forecasts, traffic density and freight movement data, these systems can even predict which sections of the road are most likely to deteriorate next. The shift from reactive to predictive maintenance will empower councils to undertake more targeted and timely interventions, preventing minor defects from escalating into much more significant hazards and costly operational disruptions. From a productivity standpoint, benefits extend across commuters, local governments and the broader economy. For commuters, smarter maintenance and monitoring translate to fewer road closures and disruptions, safer journeys and greater reliability. For councils, it enables longer-lasting infrastructure, more efficient allocation of resources and reduced budget blowouts. For the broader economy, it fosters a transport network that runs safer, smarter and stronger.
Automation in Traffic Control
Intelligent traffic management systems are also reshaping how Australia manages congestion, intersection and roadworks. Research by CSIRO demonstrates that AI engines developed for Transport for NSW can predict traffic conditions up to two hours in advance and respond within five minutes (CSIRO, 2025). Modern AI platforms integrate live camera and sensor feeds to capture detailed traffic metrics such as vehicle counts, lane usage, speed and even environmental conditions like rainfall or glare. By analysing real time data, AI systems can predict congestion patterns and dynamically optimise traffic signals (CSIRO, 2025). This has a potential in reducing wait times, rerouting vehicles, adjusting speed limits and fine-tuning commuters’ travel routes based on evolving conditions. These adaptive systems will help to achieve more traffic distribution and help alleviate peak-hour jams.
A recent project by Main Roads Western Australia (MRWA), the Planning and Transport Research Centre (PATREC), and the Pawsey Supercomputing Research Centre (PSRC) highlights how engineers are using drone-captured footage to analyse and model traffic behaviour. The collaboration involved creating advanced predictive models to design and modify intersections and roundabouts in order to ease congestion. Rafael Carvajal Cifuentes, Operational Modelling and Visualisation Manager at MRWA, describes how traditional human surveys are subject to an average 10% error margin (Pawsey, 2023). However, he says that this limitation can be overcome by automated video analysis. Drones can track vehicle speed, stopping behaviour, and gap distances with sub-second precision, to produce far more accurate datasets than manual observations. Constructing a single roundabout can cost anywhere from tens of thousands to several million dollars depending on scale and complexity. Cifuentes believes that by using AI-enhanced modelling to forecast traffic responses before construction begins, engineers can optimise designs and reduce the likelihood of costly remedial works (Pawsey, 2023).
A Growing Concern for AI’s Energy Usage
While automation is increasingly seen as a potential antidote to Australia’s productivity slowdown, it also introduces challenges of its own. The same technology that could potentially assist with efficiency across sectors is, paradoxically, resource-intensive and environmentally costly. Large language models (LLMs) and generative AI require enormous computing power and rely on vast data centres to process and store information. The training and refining of these systems demands high-performance computing on a scale that consumes significant amounts of energy (Parliamentary Business, 2025). In Australia, data centres already represent about 5% of national electricity use, with projections suggesting this figure could rise to between 8% to 15% by 2030 (Lobb, 2025). The Australia Energy Market Operator (AEMO) has begun modelling future AI-related electricity demand, and is assessing how usage can be capped (Parliamentary Business, 2025). Furthermore, AI’s environmental footprint does not stop at energy consumption, as vast quantities of water is also used to cool high-performancing computer facilities. Sydney Water, New South Wales’ largest water utility, projects that within a decade, data centres could use the equivalent of a quarter of Sydney’s annual drinking water supply. This is because cooling systems work to regulate the heat generated by thousands of powerful graphics processing units (Shine, 2025).
Clearly, AI has the potential of offering significant opportunities in enhancing productivity in the transportation and infrastructure sectors. However, Australia must work with data centres, the backbone of AI and automation, to understand the rapid evolutions of these technologies and how they can be sustainably integrated into future demand planning nationally. Doing so will be essential to building greener infrastructure and advancing low-emission transport systems that align with our long-term productivity goals. Evidently, in the race to rebuild productivity, Australia’s competitive edge will not only lie in adopting AI but deploying it wisely, efficiently and sustainably.
References
Australian Trade and Investment Commission. (2025). AWS plans to invest A$20 billion to expand digital infrastructure in Australia by 2029. Australian Trade and Investment Commission. https://international.austrade.gov.au/en/news-and-analysis/success-stories/aws-plans-to-invest-a20-billion-to-expand-digital-infrastructure-in-australia-by-2029
Bibri, S., Jagatheesaperumal, S., Ganesan, S, Jeyaraman, P. (203). Artificial Intelligence for road quality assessment in smart cities: a machine learning approach to acoustic data analysis. Springer Nature, 3 (28), 52 – 80. https://link.springer.com/article/10.1007/s43762-023-00104-y.
CSIRO. (2025). Predicting and managing traffic congestion: An artificial intelligence engine for traffic congestion management developed for Transport for NSW. CSIRO. https://www.csiro.au/en/research/technology-space/ai/Predicting-and-managing-traffic-congestion?utm_source=chatgpt.com
Field, S. (2025). Act now or fall behind: Scott Farquhar’s five-point AI plan for Australia. Forbes Australia. https://www.forbes.com.au/news/innovation/scott-farquhars-five-point-ai-plan-for-australia/
Infrastructure Australia. (2025). Infrastructure Priority List: National road maintenance backlog.Infrastructure Australia. https://www.infrastructureaustralia.gov.au/ipl/national-road-maintenance-backlog
Lobb, J. (2025). Powering Australia’s data centre boom: Navigating energy compliance and opportunity. Dentons Australia. https://www.dentons.com/en/insights/articles/2025/july/3/powering-australias-data-centre-boom#:~:text=This%20growth%20is%20energy%20intensive,%25%20under%20high%2Ddemand%20scenarios.
NEC Australia. (2025). Smarter Roads Ahead: How AI is changing the way we maintain infrastructure. NEC Australia. https://www.nec.com.au/insights/blog/smarter-roads-ahead-how-ai-changing-way-we-maintain-infrastructure
Parliamentary Business. (2025). Select Committee on Adopting Artificial Intelligence (AI) Final Report: Chapter 6 – Impacts of AI on the environment. Parliament of Australia. https://www.aph.gov.au/Parliamentary_Business/Committees/Senate/Adopting_Artificial_Intelligence_AI/AdoptingAI/Report/Chapter_6_-_Impacts_of_AI_on_the_environment
Pawsey. (2023). AI makes WA roads cheaper and safer. Pawsey. https://pawsey.org.au/case_studies/ai-makes-wa-roads-cheaper-and-safer/
Productivity Commission. (2025). Quarterly productivity bulletin – March 2025. Productivity Commission.https://www.pc.gov.au/ongoing/productivity-insights/bulletins/quarterly-bulletin-march-2025/
Shine, R. (2025). Data centres are vital for the future and AI but their environmental footprint can be a problem. ABC News Australia. https://www.abc.net.au/news/2025-08-27/ai-to-take-up-one-quarter-of-sydney-water-in-a-decade/105700928
