This page belongs to: Action Plan for Critical Technologies
Computer algorithms that automatically learn or improve using data and/or experience. Machine learning is a type of artificial intelligence. Applications for machine learning include computer vision, facial recognition, cybersecurity, media creation, virtual and augmented reality systems, media manipulation (e.g. deepfakes), content recommendation systems, and search engines
Influences all sectors of the economy, including:
- Banking & Finance
- Defence & Defence Industry
- Energy & Environment
- Transport & Logistics
- Education & Research
- Mining & Resources
Estimated impact on national interest
Economic Prosperity - High
National Security - High
Key Australian Government actions
- Digital Economy Strategy
- Artificial Intelligence Action Plan
- Modern Manufacturing Strategy
- Research Package, Centre for Augmented Reasoning at the University of Adelaide
- Next Generation Technologies Fund
- Office of Future Transport Technology – National Policy Framework for Land Transport Technology and Action Plan
- Automotive Engineering Graduate Program
- Cooperative Research Centre Projects
- AI Future Science Platform
- Defence and Strategic Goods List 2021
- More productive knowledge industries
- Fast, reliable and accurate computer vision
- Improved farm productivity, reduced reliance on chemical usage, and reduced wastage
- Improved patient outcomes from more targeted medical interventions and prescribing
- More efficient healthcare spending by reducing unnecessary treatments
- Lower rates of offending by people on bail or parole
- Less downtime for critical infrastructure from better predictive maintenance
- Reduced commute times through better traffic management
- More satisfied consumers through better product and content recommendations
- Better child development outcomes by more accurately predicting which children or families may require additional support
- Widespread application of artificial intelligence tools
- More efficient manufacturing processes
ANZ Standard Research Classification Category
- Artificial Intelligence
- Computer vision and multimedia computation
- Cybersecurity and privacy
- Data management and data science
- Graphics, augmented reality and games
- Human-centred computing
- Machine learning
- Software engineering
- Pure mathematics
Readiness Level – Now
- Product and content recommendation systems
- Fraud detection
- Creating highly realistic deepfakes and synthetic media
- Spam filtering
- Detecting potential malware or cyber intrusions
- Detecting harmful or unlawful content
- Public transport planning
- Precision agriculture
- Predictive and smart policing algorithms
Readiness Level – 2–5 years
- Computer vision for autonomous vehicles
- Reliably detecting deepfakes and synthetic media
- High-resolution weather models
- Creating high quality synthetic datasets to train other machine learning models
- Creation of voice and computer generated actors and scenes
Readiness Level – Beyond 5 years
- Fully autonomous systems – driverless vehicles and drones
- Personalised medicine
- Responding to novel cyber-attacks
- Creation of novel pharmaceuticals
- Personal assistant and companion robots
Australia's place in the world
Australia ranks 9th globally for research impact, led by the University of Melbourne. The United States dominates research on machine learning, and also has the top five international institutions. Overall publications have increasing at 22% p.a.
Australia ranks 13th for venture capital (VC) investment, which is led by the United States and China. VC investment for machine learning has been increasing at 18% p.a. since 2016. Patents have been increasing at 59% p.a. since 2015, led by China, ahead of the United States, with Australia in 17th.
Opportunities and risks
Machine learning has significant potential to affect Australia’s economic prosperity and Australia’s national security.
Opportunities for machine learning focus on increasing productivity and making artificial intelligence use cases more independent, practical and economical. Machine learning can automate or streamline many monotonous, time-consuming or low-value activities, freeing people to focus on more valuable or less unpleasant tasks. It has been estimated that up to 46% of jobs may be automated through automation and existing workers will be financially better off by between $4,000 to $15,000 p.a. by 2030.
Some of the key risks from machine learning are changing workforce requirements, poorly designed use cases, overly-optimistic reliance on machine learning‑based tools, and intentional use of machine learning-based tools to undermine Australia’s national interest. Falling costs may mean that businesses automate some roles and reduce staffing levels; this has the potential to create new and different opportunities for the workforce. Machine learning to automate processes and functions is not a one-size-fits-all solution and there are risks associated with applying the technology to situations before it is ready or without managing limitations during development and implementation. These risks range from lost money or opportunities, up to and including physical harm to people and property.
Machine learning also has the potential to underpin new cyber or physical weapons that could be used by or against Australia. To combat this, awareness and technical understanding of machine learning technology will be vital.
Research impact (RI)
The United States has the highest research impact in this field, with Australia ranked 9th globally. Total volume of published research has increased at around 22% p.a. over the 5 year period 2016–2020, with 21% of all research involving international collaboration.
- USA - 65934
- China - 30488
- UK - 23109
- Germany - 9126
- Canada - 8678
- Australia - 5518
The research impact provides an indication of the productivity of a country or institution. Here, productivity was assumed to be represented by the volume of publications (i.e. scholarly output) as an indicator of the resources and facilities, and the level of interest in the publications as an indicator of quality.
Australia ranks 14th for venture capital (VC) investment in machine learning, with the United States (1st) and China (2nd) leading VC investment ahead of the United Kingdom, Canada and Israel. Investment in this area has been growing at 18% p.a. since 2016.
Data from Crunchbase. The Crunchbase database provides a partial view of the global VC landscape. However the quantity, quality and richness of the data are considered to be statistically significant, and indicative of global trends.
Patents - international
Most patents for this technology were filed by Chinese applicants or inventors, ahead of the United States. Australia ranks 17th. Global patent applications have been increasing at 59% annually since 2015.
- China - 29781
- USA - 21038
- R. of Korea - 4613
- Japan - 3968
- Taiwan - 2957
- Australia - 321
Research institutions - international
The United States has the top five international institutions, with Alphabet Inc. (a.k.a. Google) having the highest research impact. The rest of the top 10 institutions are from the United Kingdom (3) and China (2).
|Rank||Top International Institution||Research Impact|
|1||Alphabet Inc. | United States||10284|
|2||Stanford University | United States||5929|
|3||Massachusetts Institute of Technology | United States||4534|
|4||University of Washington | United States||4398|
|5||Harvard University | United States||4167|
|6||University of Oxford | United Kingdom||2974|
|7||Imperial College London | United Kingdom||2893|
|8||Chinese Academy of Sciences | China||2672|
|9||University College London | United Kingdom||2567|
|10||Tsinghua University | China||2507|
Research Australian - Australia
Within Australia, the University of Melbourne and the University of Technology Sydney have the highest research impact.
There are no Australian institutes in the top 50 international institutions.
|Rank||Top Australian Institution||Research Impact|
|1||University of Melbourne||590|
|2||University of Technology Sydney||563|
|4||University of Sydney||371|
|5||University of Adelaide||354|
|8||University of New South Wales||284|
|9||University of Queensland||274|
Patents - Australia
|Top 5 Australian Patent Applicants||Patent Families|
|n/a (private citizen applicants only)||3|
|University of Melbourne||2|
|Swinburne University of Technology||2|
Patents filed by Australian businesses, 2015–2019.