Artificial intelligence (AI) technologies can solve problems on their own without human intervention. AI could improve production efficiency, safety and quality in almost every industry.
Example technologies
- Machine learning, including neural networks and deep learning
- AI algorithms and hardware accelerators
- Natural language processing, including speech and text recognition, analysis and generation
Example applications
These technologies can be used for:
- facial recognition and detection
- automating manual processes
- virtual assistants
- advanced cyber security
- automatically classifying objects in images
- reliable self-driving vehicles
- neuromorphic computing (computers modelled on the human brain and nervous system)
- rapid data analysis
- better traffic management
- automated investing (using computer algorithms to generate tailored financial advice for people).
Global research trends
Research into AI has been growing steadily over the past 10 years, with over 180,000 AI research publications published around the world in 2021.
Number of AI research publications published globally each year, 2013–2021
Global research rankings
China and the United States do the most AI research.
The graph below shows the number of AI publications each jurisdiction published between 2018 and 2022. It also shows what proportion of each jurisdiction’s publications were ranked in the top 10% of publications worldwide. This is based on how often the jurisdiction’s publications on core AI research subjects were cited.
AI publications by jurisdiction, 2018–22
Show text version
|
Ranking |
Jurisdiction |
Publications, 2018 to 2022 |
Percentage in top 10% |
|
1 |
China |
418,330 |
16.2% |
|
2 |
USA |
296,329 |
21% |
|
3 |
India |
118,848 |
9.4% |
|
4 |
UK |
88,402 |
18.2% |
|
5 |
Germany |
86,090 |
16% |
|
6 |
France |
65,597 |
12% |
|
7 |
Italy |
60,012 |
13.9% |
|
8 |
Canada |
57,101 |
18.6% |
|
9 |
Japan |
56,571 |
9.8% |
|
10 |
South Korea |
49,266 |
17.3% |
|
12 |
Australia |
46,435 |
22.0% |
Commercialisation
Patent filings are a way to measure how innovation is being commercialised. Patent data can also be used to identify potential collaborators and export markets.
The graphs below show numbers of patent families filed. A patent family is a set of patents filed in different jurisdictions for the same invention.
AI patent families filed globally each year, 2017–2021
Show text version
|
Year |
Global patent families filed |
|---|---|
|
2017 |
18,382 |
|
2018 |
32,752 |
|
2019 |
52,478 |
|
2020 |
74,932 |
|
2021 |
65,771 |
Jurisdiction of origin for AI patent families filed, 2017–2021
Show text version
|
Ranking |
Jurisdiction |
Global patent families filed |
|
1 |
China |
174,154 |
|
2 |
United States |
65,188 |
|
3 |
South Korea |
17,058 |
|
4 |
Taiwan |
16,090 |
|
5 |
Japan |
11,524 |
|
13 |
Australia |
1,459 |
The Australian industry
Australian universities, research organisations and companies do world-leading AI research.
We have an opportunity to further coordinate and concentrate our research capabilities on nationally significant matters, including creating AI industries that generate jobs.
AI research output (number of publications) by state and territory, 2018–2022
Show text version
- NSW: 35%
- Victoria: 28%
- Queensland: 14%
- ACT: 8%
- Western Australia: 6%
- South Australia: 7%
- Tasmania: 1%
- Northern Territory: 1%
Collaboration
The jurisdictions we collaborate with most on AI research are China, the United States, the United Kingdom, India and Iran.
Australia’s top 5 global collaborators on AI research, 2018–2022
Show text version
- China: 7,487 publications
- United States: 3,020 publications
- United Kingdom: 1,695 publications
- India: 1,029 publications
- Iran: 795 publications
Australia wants to be a leader in developing international technical standards for AI. We are supporting Australian industry to engage more effectively with standards development organisations.
The future
Australia must continue to improve its AI capability so we can:
- become a world-leading digital economy
- enable more inclusive job opportunities
- help industry and governments reduce costs, improve efficiency and productivity, and be more competitive
- use data to make better decisions
- counter national security threats and improve our nation’s defence with smarter military systems and operations.
Australia is focusing on 3 areas of AI to transform existing industries and build new ones:
- AI for better health, aged care and disability services
- AI for better towns, cities and infrastructure
- AI for better natural resource management.
Market opportunities
- The CSIRO estimates that developing and commercialising AI will add over $22 trillion to the global economy by 2030.
- The CSIRO estimates that digital innovations including AI could contribute $315 billion to Australia’s GDP by 2030.
- Australian spending on AI systems will grow to over $3.6 billion by 2025, at a compounding annual growth rate of 24.4% between 2020 and 2025.
Australian businesses have reported a range of benefits from using AI, including:
- an average revenue benefit of $361,315
- time savings of around 30% across all AI initiatives they have implemented
- an increased market demand to accelerate and expand their AI-solutions, while reducing business costs.
Data sources
Publication and citation data on this page was collected from Clarivate’s Web of Science and InCites. Data analysis was performed by CSIRO.
Patent data collection and analysis was performed by IP Australia using data from the European Patent Office’s PATSTAT 2022 autumn edition.