Australia’s artificial 
intelligence ecosystem

Catalysing an AI industry

December 2023

CSIRO Australia’s National Science Agency



Citation

Hajkowicz S, Bratanova A, Schleiger E, Naughtin C (2023) 
Australia’s artificial intelligence ecosystem: Catalysing an 
AI industry. CSIRO, Canberra.

The authors would like to thank the many experts and 
stakeholders we interviewed from Australia’s artificial 
intelligence companies and research institutes. We are also 
grateful to staff within CSIRO who provided review and 
feedback on early draft versions of this report.

Copyright

The contents of this report are subject to the 
Creative Commons licence called “Attribution-Non 
Commercial‑Share Alike 4.0 International (CC BY-NC-SA 4.0)”.

Disclaimer

CSIRO advises that the information contained in this 
publication comprises general statements based on 
scientific research. The reader is advised and needs to 
be aware that such information may be incomplete or 
unable to be used in any specific situation. No reliance 
or actions must therefore be made on that information 
without seeking prior expert professional, scientific and 
technical advice. To the extent permitted by law, CSIRO 
(including its employees and consultants) excludes all 
liability to any person for any consequences, including 
but not limited to all losses, damages, costs, expenses 
and any other compensation, arising directly or indirectly 
from using this publication (in part or in whole) and any 
information or material contained in it.

Acknowledgements

The National AI Centre is funded by the Australian 
Government and coordinated by CSIRO.

CSIRO acknowledges CEDA and Google as 
Foundation Partners of the National AI Centre

About this report

This report was prepared by CSIRO for the National 
Artificial Intelligence Centre. It presents information 
about companies and research institutes active in 
developing and applying artificial intelligence in Australia. 
This provides investors, customers, policy-makers, 
workers and students with information about Australia’s 
capabilities, service offerings and research specialisations 
in artificial intelligence.

Accessibility

CSIRO is committed to providing web accessible content 
wherever possible. If you are having difficulties with 
accessing this document please contact csiro.au/contact.



Contents

Executive summary...................................................................................................................................2Introduction.................................................................................................................................................6Section 1: Artificial intelligence companies...................................................................................8How we identify and define an AI company.....................................................................................................................9Product and service offerings.........................................................................................................................................11Geographic patterns of AI companies.............................................................................................................................15Company formation and maturity...................................................................................................................................22Jobs and skills demand.....................................................................................................................................................24Investment trends.............................................................................................................................................................25Section 2: Research and development.........................................................................................26Methods and source data................................................................................................................................................27Knowledge creation (research publishing trends).........................................................................................................29Application domain specialisations................................................................................................................................31Artificial intelligence technology specialisations..........................................................................................................33Universities and research institutes................................................................................................................................35Product innovation and patent applications..................................................................................................................38Section 3: Industry stakeholder and expert perspectives...................................................42Insights from our interviews............................................................................................................................................43Australia’s AI ecosystem case studies..............................................................................................................................47Key takeaways...................................................................................................................................................................52Planned improvements....................................................................................................................................................54References..................................................................................................................................................55

Executive summary

Australia has a growing and diversified AI industry

We identified 544 companies in Australia whose main 
business activity is developing and selling AI products and 
services. These companies are nestled within a broader 
ecosystem of technology companies with >336,000 firms 
in the professional, scientific and technical services 
industry and close to 25,000 firms in information media 
and telecommunications.

Our initial search on company databases returned 
1,606 AI companies. This was further screened, verified and 
validated down to 544 active AI companies. We developed 
a taxonomy to assign 10 AI product and service offering 
categories to each company as follows.

Why do they add up 
to >544 companies?

It’s because a single company 
can have multiple product 
and service offerings.

The number of AI companies 
in Australia has increased 
substantially over the past decade

396

Number of Australian AI 
companies opened for 
business in the past 10 years

204

Number of Australian AI 
companies opened for 
business in the past 5 years

7.7%

Year-on-year growth of company 
count over the past 5 years

Australia’s AI companies are young and nimble

6 yrs

Median age of 
Australian AI companies

13 yrs

Median age of Australian information and 
communications technology (ICT) sector companies

Australia’s AI companies 
are showing patterns of 
geographic clustering

8

Number of tightly clustered hotspots (precincts) 
of AI companies across parts of Sydney, 
Melbourne, Brisbane, Adelaide and Perth

54%

Share of Australia’s 544 AI companies 
located in these 8 clusters 
(296 companies in total)

80 m

Average straight-line distance between 
an AI company and its nearest 
neighbouring AI company within 
the same cluster



Australia’s share of AI publishing and product innovation

1.2%

Australia’s share of global 
publishing on all topics

1.6%

Australia’s share of global 
publishing on AI topics

0.2%

Share of global AI patent applications 
with an Australian inventor

Australia’s top 3 AI application domains

Australia’s R&D sector (public and private) is publishing research 
in AI application domains at a faster rate than the global average 
in these areas. The revealed technology advantage (RTA) scores 
show the rate at which we’re publishing, where a value of 
1 would be equal to the global rate.

4.8x

AI applications in 
livestock production

3.8x

AI applications in medical 
laboratory technologies

3.2x

AI applications 
in horticulture

Australia’s top 3 AI technology sub‑fields

Australia’s R&D sector (public and private) is publishing research 
in AI technology types at a higher rate than the global rate in 
these areas. The RTA scores show we’re publishing at 7.6 times 
the global average rate in the field of self-organising maps 
(a type of neural network).

7.6x

Self-organising map 
(a type of neural network)

5.3x

Ant colony optimisation (an algorithm 
inspired by ant colony behaviours)

5.2x

Gradient tree boosting (a machine 
learning technique used for prediction)

Australia’s top 6 universities and research institutes 
by AI publishing intensity

The share of AI research published by Australian universities is rising, with >10% of 
the research published from the following six institutions being on AI topics in 2021.

18.8%

University of 
Technology 
Sydney

13.3%

University 
of Southern 
Queensland

12.3%

CSIRO

12.2%

Swinburne 
University of 
Technology

10.1%

University of 
New South 
Wales

12.9%

Deakin
University



Stakeholder perspectives

We held interviews with 28 stakeholders across Australia’s 
AI ecosystem including AI companies (startups and small, 
medium and large enterprises), academia and government 
agencies. These consultations explored current and 
emerging areas of competitive advantage across the 
Australian AI ecosystem, the unique value proposition 
of Australian‑developed AI and opportunities to support 
the ecosystem in the future.

Challenge in separating the hype from reality. 
Leaps forward by open-access tools like ChatGPT 
have accelerated interest in AI technologies. Business 
leaders can better visualise the potential benefits AI could 
bring to their organisations, but many remain risk-averse 
and apprehensive about AI technologies.

Need more awareness of local AI service 
providers. Australian companies are opting for 
international AI service providers when the capabilities 
do not exist or cannot be identified locally. There are 
opportunities to expand platforms like the Australian AI 
Ecosystem Discoverability Platform to provide greater 
coverage and promote awareness of Australia’s existing 
AI ecosystem.

Deciding on what AI to build, buy or borrow. 
To be globally competitive, Australia needs to set 
conditions that support AI adoption as well as create 
unique AI capabilities. This could include addressing the 
persisting challenge of scaling Australia’s strong academic 
and research outputs into commercial innovations.

Prioritising being an AI specialist over an AI 
generalist. Acknowledging our population, market 
and funding size can limit Australia’s capacity to compete 
in the global AI arena, there could be opportunities to 
grow world‑leading capabilities in strategic areas of AI 
specialisation. These areas of AI specialisation should be 
based on domains where Australia has a strategic and/or 
comparative advantage.



An opportunity for safe and responsible AI. 
There is a significant opportunity for Australia to 
be a global leader in the development and use of safe and 
responsible AI technologies. The ‘Australian brand’ is a 
trusted asset, which puts us in a strong position to drive 
future safe and responsible AI developments.

The benefits of socio-cultural diversity. 
Australia’s cultural and linguistic diversity provides 
a unique comparative advantage when it comes 
to developing novel AI technologies. Australia is 
well‑placed to leverage this diversity in the creation 
of novel AI technologies, as well as position itself as 
an AI technology testbed.

Strengthening linkages across the AI ecosystem. 
Australia’s AI ecosystem requires strong leadership 
and improved connectivity across the sector. Doing so 
will maximise opportunities for Australia to define and 
strengthen its global reputation in AI and mature as 
an ecosystem.

Growing Australia’s AI talent and business 
ecosystem. Talent shortages are a persisting 
struggle reported by Australian businesses, particularly 
when it comes to sourcing specialised technical skills. 
Australia could consider new approaches to inject talent 
into the AI ecosystem and reduce friction for professionals 
working in these areas.



Introduction

What is this report about?

This is the second report about Australia’s artificial 
intelligence (AI) ecosystem with the first being published 
in 2022 (NAIC, 2022). It has been commissioned by the 
National Artificial Intelligence Centre (NAIC) and delivered 
by CSIRO. Australia’s AI ecosystem is comprised of startups, 
small-to-large-sized companies, universities, research 
institutes, industry organisations and public sector agencies 
engaged in developing and applying AI technologies. 
This report aims to provide a snapshot of the current state 
of Australia’s AI ecosystem and inform future strategy and 
policy decisions about its growth and development.

Why read this report?

This report contains information about companies, product 
innovation, investment, universities, research institutes and 
specialised capabilities relating to AI in Australia. It also 
contains information about the perceived challenges and 
opportunities for future growth and development of AI 
in Australia. The report aims to inform people in industry, 
academia, government and community spheres making 
decisions about AI capability uplift.

This report is useful to anyone wanting to understand 
and develop AI capability in Australia. It could be used 
by investors looking for opportunities in the Australian 
marketplace. It could be used by a company wanting to 
train and recruit AI talent. It could be used by someone 
starting or developing an AI company. It could be used 
by universities and research institutes planning for 
AI capability development. It could be used by Local, 
State/Territory and Federal government agencies aiming 
to support the AI industry and job growth. And it could be 
used by workers and students planning their AI careers.

What is the Australian 
AI ecosystem?

The AI ecosystem comprises companies (startups, small, 
medium and large enterprises), universities, education and 
training institutes, research institutes, industry bodies and 
professional associations concerned with the development 
and application of AI in Australia. Whilst practically all 
organisations use AI to some extent, this report focuses 
on companies and research institutes that have a primary 
or substantial focus on AI. We identify several sub-sectors 
within the AI ecosystem:

1. Small, medium, large and startup companies that 
make and sell AI products and services
2. Research institutes and universities significantly 
engaged in developing and applying AI
3. Venture capital and other investors who inject 
capital into the AI industry
4. Industry associations, professional bodies 
and community organisations
5. Education and training institutions1
6. Government bodies that support and regulate 
the AI industry1
7. Workers in AI occupations1





1 These sub-sectors were out of scope for the current report as we currently do not have datasets or standardised reporting frameworks for capturing 
information on these sub-sectors. We hope that future versions of this report will be able to shed light on these other parts of the Australian AI ecosystem.



Why is Australia’s AI 
ecosystem significant?

Australia, along with most advanced economies, 
is grappling with an ongoing productivity slump 
(Australian Government, 2023). This causes lower rates 
of economic growth and a reduced growth rate in living 
standards. AI is a general-purpose technology that could 
turn things around. AI is impacting every policy sphere, 
industry sector and (practically) every career. This impact 
will intensify in the future.

When used well, AI allows people to do things faster, 
cheaper, safer and better. Just like electricity in the 
1920s, AI has the capability to deliver an economy-wide 
productivity boost in the 2020s and beyond. Many new 
jobs, startups and corporations will be created through 
AI over the coming years. Many existing companies and 
workers will be impacted, which is part of the reason 
why AI is attracting so much attention. It is both a critical 
enabler of economic activity and a driver of disruption.

What’s changing in the world of AI?

The year 2023 is considered by some as a new phase in 
the journey of AI. We’ve seen the rapid ascendancy of 
foundation AI models and generative AI such as ChatGPT, 
DALL-E 2 (OpenAI), GitHub Copilot, Imogen, Codey, Chirp 
(Google), Llama 2 (Meta AI) and others. Many more such 
models are likely to emerge over the coming months and 
years. Foundation models such as these are typically trained 
on vast quantities of data and can be used to perform a 
wide variety of tasks and solve problems across virtually 
all industries.

The global AI sector is racing to build new foundation 
models and plug-ins to existing models. This will be 
associated with the emergence and rapid scale-up of new AI 
companies. It will also change what’s possible for existing 
companies with the prospect of a significant productivity 
uplift. There is much uncertainty about how the next 
generation of foundation AI models will play out, but a 
significant impact on Australia’s economy and industry is 
certain. It’s already happening.

Where does our information 
come from?

To describe the AI ecosystem, we draw upon large 
commercial databases which are often used by investors 
when searching for companies such as Crunchbase and 
PitchBook. These databases contain detailed industry 
classifications and wide-ranging company data. Our 
analysis of AI product innovation and research activity 
largely comes from The Lens, a comprehensive database 
on Australian and global scholarly publishing and patent 
applications. We also complemented these analyses 
with structured qualitative interviews with experts 
and stakeholders across the Australian AI ecosystem.

What’s in this report?

The report is broken up into three main sections. The first 
section presents information on the AI industry in Australia 
through the lens of startups and small, medium and large 
companies whose primary business model involves making 
and selling AI products and services. We examine the 
temporal and spatial patterns of these AI companies and 
classify their AI offerings. The next section focuses on the 
research and development sector, examining patterns of 
AI research publishing and patent applications in Australia 
and areas of global specialisation. We also present 
information about the AI research activities of Australia’s 
leading universities and research institutes. The final 
section presents insights from interviews with experts 
across Australia’s AI ecosystem, shedding light on the 
opportunities and barriers facing the ecosystem.



 1 
Artificial intelligence 
companies

Startups and small, medium and large 
Australian‑headquartered companies 
whose primary business activity is making 
and selling AI products and services to customers.



How we identify and define 
an AI company

We define an Australian AI company as a privately owned or publicly listed company 
headquartered in Australia whose primary business activities involve making and selling 
AI products and services to customers. The list of Australian AI companies was constructed 
using three datasets: Crunchbase (Crunchbase, 2023a), Pitchbook (Pitchbook, 2023) and the 
Australian AI Ecosystem Discoverability Platform Beta (CSIRO, 2023).

Crunchbase

Over 2 million companies’ 
data including over 81,000 
Australian companies

703

organisations 
discovered

OECD search phrases recommended 
for AI companies, goods and services 
search: 105 keywords and phrases

831

organisations 
discovered

Validation

Cross checking
of lists

Verification

Manual scan 
of company 
websites

PitchBook

Over 3.6 million companies’ data 
including over 40,000 companies 
with offices in Australia

Australian AI Ecosystem 
Discoverability 
Platform Beta

NAIC directory of Australian capabilities 
to enable business adoption of AI 
(126 companies) and NAIC company 
list (219 companies)

345

organisations 
recorded

544

Australian AI
companies

531 privately owned 
organisations

13 publicly listed 
organisations

Headquartered in Australia

Making and selling AI 
products and services



We applied a set of 105 AI keywords developed by the OECD 
(Nakazato and Squicciarini, 2021) to search the descriptions 
of companies headquartered in Australia. Our aim was to 
identify companies whose main product or service offering 
was about AI. We recognise that many (arguably most) 
companies use AI; but in this analysis, we’re after companies 
who make and sell AI products or services as their main 
business activity.

As a result of the search through three databases 
(Crunchbase, Pitchbook, NAIC AI Ecosystem Discoverability 
Platform Beta and company list), we identified a list of 
1,606 Australian companies with unique names. The list 
of companies was further validated and verified by 
cross‑checking datasets and manual screening of company 
web pages and LinkedIn profiles. Additional companies 
identified through the stakeholder consultations were also 
included in the final set, which totalled 544 Australian AI 
companies, including 531 privately owned companies and 
13 publicly listed companies.

To put these findings in perspective, Australia is home 
to over 360,000 companies operating in the industries 
where most AI companies operate; namely, professional, 
scientific and technical services and information media and 
telecommunications. According to the Australian Bureau 
of Statistics (2023), at the end of June 2023, there were:

• 336,214 businesses in Australia in the professional, 
scientific and technical services industry. The industry 
demonstrated the third largest net increase in business 
count in the past year (an increase of 4,960 businesses) 
after healthcare and social assistance (10,721) and 
transport, postal and warehousing (5,040)
• 24,757 businesses in the information media and 
telecommunications industry in Australia. Compared 
with the year before, the number of businesses in the 
industry dropped by 0.3% or 75 companies.


Limitations of our data 
and methods

Identifying AI companies is a challenging task due to the 
nature of AI as a general-purpose technology and the 
range of ways it is used across all sectors of the economy. 
Technology companies also make a large and dynamic 
marketplace with company entries and exits occurring 
frequently. The following limitations of this analysis need 
to be considered when interpreting and using the results:

• We do not include overseas technology corporations 
with a presence in Australia because we cannot measure 
the nature and extent of their Australian (versus global) 
activity. However, they are likely to represent 
a significant share of Australia’s AI activity.
• While best efforts were made to collect data on as 
many Australian AI companies as possible, the list is 
not exhaustive. The companies with more general 
descriptions of their activities on Crunchbase and 
Pitchbook platforms could have been overlooked 
in our search.
• The coverage of the datasets used (Crunchbase, 
Pitchbook) has been expanding over time, so trend 
data should be interpreted with caution.
• Data availability on the operations of private companies 
and publicly listed companies varies, so any comparison 
between these two groups should be performed 
with caution.
• The list of AI product and service offerings is sourced 
from NAIC and complemented by the analysis of 
Crunchbase industry classifications. This list is not 
exhaustive and the classification categories are not 
mutually exclusive.
• For industry analysis, we applied the Crunchbase industry 
taxonomy, which differs from the taxonomy used by the 
Australian Bureau of Statistics.


Overall, the number of AI companies is likely to change 
over time. Companies will grow, merge, enter and exit. 
Definitions of AI will change as the technology changes. 
Therefore, our estimate of AI companies in this report is 
a snapshot for the time of publication.



Product and service offerings

Working closely with the NAIC, we identified a set of product and service offering categories for Australian AI companies, 
noting that a company might provide multiple types of offerings. The categories are as follows:

Cloud services

Cloud services include software (an app or set of apps), 
application development platforms and infrastructure 
capabilities (memory, storage, networking). Cloud services 
enable businesses to develop and implement AI capabilities. 
Examples of AI cloud services capabilities include 
development and implementation for SaaS (Software as a 
Service), PaaS (Platform as a Service), IaaS (Infrastructure as 
a Service).

Consulting

Consulting services assist with strategy, organisational 
change processes and methods for optimal utilisation and 
application of AI technologies. Consulting services enable 
AI adoption and commercialisation and include services in 
innovation consultancy, legal and financial services.

Finished solutions

Finished solutions providers create ready to use AI 
applications, including platforms, enabling businesses 
to adopt and implement AI technologies. Finished 
solutions providers offer AI services to companies to help 
solve business challenges, including end-to-end design, 
implementation and maintenance of AI technologies for 
specific business use cases.

Hardware and infrastructure

Hardware and infrastructure vendors provide physical 
and network components of AI technologies to enable 
computational capabilities and automation. Key 
infrastructure for AI systems is high computing capacity 
(CPU, GPU-based computing), storage capacity, networking 
infrastructure (high bandwidth, low latency) and security. 
This group of AI capability providers also include developers 
of robotic solutions, communications and data centres 
infrastructure, applications of sensor technology.

Systems integration

Systems integration providers develop solutions that 
combine different computing operations components, 
including hardware and software for a more coherent 
and capable system. This includes developing software 
components, such as speech synthesisers and 
common‑sense knowledgebases, that are interoperable 
with other components, and planning, coordinating, 
scheduling, testing and improving computing operations.

Data service

Data services provide resources and expertise for the 
utilisation of data to develop and optimise AI capabilities, 
including datasets, data analysis and management. 
This capability includes providers of DaaS (Data as a 
Service), and web-delivered services that perform various 
functions on data, such as data mining, visualisation, 
image and speech recognition and others.

Business transformation

The business transformation AI capability includes 
technology solutions and services in digitalisation, 
digital transformation and broader business integration. 
This category includes providers of business and 
management information systems, and contract and 
document management solutions.

Research and innovation

Research and innovation services provide expertise and 
technical capabilities to enable cutting-edge progress on 
emerging AI domains and solutions to complex challenges. 
This category encompasses organisations seeking to 
push the boundaries of knowledge and explore what is 
possible in terms of AI applications, including the pursuit 
of IP that can result in new products and services for 
commercialisation.

Governance and ethics

The governance and ethics capability group includes 
providers of advisory services that focus on the responsible 
and ethical adoption of AI technologies in accordance 
with ethical, privacy and security considerations and 
requirements. Companies promoting inclusion and 
diversity in the industry are also classified under this 
capability group.

Skills and training

Skills and training providers enable individuals and 
organisations to learn and build new workforce capabilities. 
This category includes companies that provide online 
training and courses along with tailored solutions for 
workforce adoption and upskilling.



Method for classifying companies

The table below shows the correspondences for the NAIC AI ecosystem taxonomy, Crunchbase Industry List and company 
descriptions keyword search.

AI ECOSYSTEM 
CAPABILITY 
CATEGORY

KEYWORDS

CORRESPONDING INDUSTRIES FROM 
CRUNCHBASE INDUSTRY LIST

Cloud services

Cloud technolog*, Cloud Comput*, Cloud Infrastructure, 
Platform Service, Platform Vendor, Software as a Service, 
SaaS, Platform as a Service, PaaS, Infrastructure as a 
Service, IaaS, cloud service*

Cloud Infrastructure, Private Cloud, Cloud Data 
Services, Cloud Management, Cloud Security, 
Cloud Security, Cloud Computing, SaaS

Finished solutions

AI application*, AI as a Service, Artificial Intelligence as a 
Service, Artificial Intelligence Development Service*, AI 
Development Service*, Artificial Intelligence Solution*, AI 
solution*, App*, platform*, game*, AI model*, software* 

App Discovery, App, Consumer Application, 
Enterprise Application, Mobile App, Web 
App, Software 

Hardware and 
infrastructure

AI Infrastructure, AI Hardware, Artificial Intelligence 
Infrastructure, Artificial Intelligence Hardware, GPU, CPU, 
GPU-based computing, networking infrastructure, FPGA*, 
on-chip memory, SRAM, automat*, robot*, sensor*

Application Specific Integrated Circuit, ASIC, 
Augmented Reality, Communication Hardware, 
Communications Infrastructure, Computer Vision, 
Data Center, Data Center Automation, GPS, GPU, 
Network Hardware, Sensor, Virtual Reality, IT 
Infrastructure, Network Security

Systems integration

Systems integrat*, speech synthesizer*, common sense 
knowledgebase*, AI system*, Artificial Intelligence 
system*, integrat*, design architect*, chatbot*, symbiosis, 
AI standard*

Data Integration

Business 
transformation

Business transform*, digital* transform*, digitali*, optim* 
process, business solution*, solv* business, digital 
product*, management tool*, business system*, growing 
business*, digital ecosystem, business intelligence

Business Information Systems, Contact 
Management, Document Management, 
Management Information Systems

Data services

data analy*, data management, Data as a service, DaaS, 
Training data, data aggregat*, data annotation, data 
architecture, data transit, data storage, data analysis, 
visuali*, voice*, speech*, vision*, data analy*, big data*, 
video analy*, conversation*, art*, music*, image recogni*, 
knowledge graph*, simulat*, interpret*

Application Performance Management, Business 
Intelligence, Data Mining, Data Visualization, 
Facial Recognition, Geospatial, Image Recognition, 
Intelligent Systems, Location Based Services, 
Predictive Analytics, Quantified Self, Speech 
Recognition, Text Analytics, Usability Testing, 
Big Data, Analytics

Consulting

Consulting Service*, Consult*, Innovation Consult*, 
innovation service, scal* up, legal*, decision*, insurance*, 
financial report*, report* servic*, financial servic*, 
business advis*, data-driven*, marketing

Consulting, Financial Services, FinTech, Trading 
Platform, Market Research, Advertising, Marketing

Research and 
innovation

research*, Innovation servic*, Artificial Intelligence 
research, AI research, knowledge broker*, research 
broker*, science, scientist*

Governance 
and ethics

Governance, ethic*, privacy, female*, woman*, 
indigenous,*securit*, Data govern*, data privacy, cyber*

GovTech, Information Services, Cyber Security, 
Government, Military, National Security, 
Network Security

Skills and training

skills, training*, workforce*, learning, upskill, workforce 
capabilit*, educat*, employee assist*, recruit*

Continuing Education, Corporate Training, 
E-Learning, EdTech, Education, Edutainment, 
Higher Education, MOOC, Personal Development, 
Primary Education, Secondary Education, Skill 
Assessment, STEM Education, Training, Tutoring, 
Vocational Education





*Used as a common wildcard symbol to broaden the search to include the words that start with the same letters.



Product and service offerings of Australian AI companies

To understand the capabilities of the Australian AI 
ecosystem, we applied the NAIC taxonomy of AI 
product and service offering categories to the company 
descriptions. We incorporated the results of the NAIC 
survey (CSIRO, 2023) and the Crunchbase industry group 
classifications (Crunchbase, 2023b) to better define what 
capabilities each company contributes to the Australian 
AI ecosystem. Each Australian AI company was classified 
into at least one of the ten AI product and service 
offering categories.

Of the 544 Australian AI companies identified in this 
analysis, 395 (73%) provide capabilities in data services. 
Data services include expertise for the utilisation of data to 
develop and optimise AI capabilities, including data analysis 
and management. This component of AI capability includes 
providers of DaaS, data mining, visualisation and image and 
speech recognition, among others.

A large share of the companies (65%, 356 companies) are 
also classified as providers of finished solutions, which 
include ready to use AI applications (e.g., platforms) that 
enable businesses to adopt and implement AI technologies. 
Finished solutions providers also offer AI services to 
companies to help solve business challenges, including 
end-to-end design, implementation and maintenance of 
AI technologies for specific business use cases.

Roughly every two out of five AI companies are engaged 
in consulting and over 200 companies provide capabilities 
in skills and training. Skills and training providers enable 
individuals and organisations to learn and upskill for 
new workforce capabilities in AI-related domains. 
Another leading share of AI capabilities is in the domain of 
AI governance and ethics (19%), which includes companies 
actively promoting diversity and inclusion in the industry.

Capabilities each company contributes to the Australian AI ecosystem



AI companies by industry grouping

We used industry groupings available in commercial 
databases to map AI companies by industry group. 
Based on Crunchbase industry group taxonomy, 88% 
of Australian AI companies are classified as software 
providers and 79% of companies work in data and analytics. 
This result is not surprising given the nature of AI 
development and adoption. Importantly, 62% of AI 
companies are also in the science and engineering domain, 
which might reflect innovation, research and development 
activities. Beyond this, the largest share of AI companies 
are working in information technology (IT), professional 
services, financial services and healthcare.

Top 20 industry groups for Australian AI companies (n = 461)

Note: A total of 461 Australian AI companies had available data on industry group classification and were included in this analysis. One company can have 
multiple industry groups, so it would be counted across multiple industry groups.



Geographic patterns 
of AI companies

We mapped the locations of AI company headquarters across Australia. Australian AI 
companies are located in all States and Territories, with the majority in New South Wales, 
Victoria and Queensland. The following heat maps show the intensity of clustering. 
While the most intense clusters are in Australia’s capital cities, regional areas also feature.

Locations of AI company headquarters across Australia

This map was created using the Google Geocoding API and data from Crunchbase on 
the AI company name, short description, web address and general location (city/suburb and 
State/Territory). We manually checked and edited the company addresses with web searches. 
With the point (headquarters) data we used the “Folium HeatMap” Python plugin which uses 
a density estimation algorithm to create a heatmap.

Map tiles by Stamen Design, under CC BY 3.0. Data by © OpenStreetMap, under CC BY SA.

Clustering intensity

LOW

HIGH



While most Australian AI companies are located in 
greater capital cities areas, 29 are in regional areas.

AI companies by location

AI companies by greater capital city

AI companies by regional areas

NUMBER OF AUSTRALIAN AI COMPANIES








	



Rest of TasmaniaGreater HobartGreater DarwinRest of VictoriaRest of New South WalesRest of QueenslandGreater CanberraGreater AdelaideGreater PerthGreater BrisbaneGreater MelbourneGreater Sydney

AI company heatmap – Greater Sydney

AI company heatmap – Greater Melbourne

CITY CENTRE

LOW

CITY CENTRE

LOW



AI company heatmap – Greater Brisbane

AI company heatmap – Greater Perth

CITY CENTRE

LOW

CITY CENTRE

LOW



AI company heatmap – Greater Adelaide

Clustering intensity of AI companies in Australian cities – Nearest Neighbour Analysis

Clusters are hotspots of concentrated industry activity and have been shown to be important in the formation, growth 
and development of industries (Hajkowicz et al., 2023a; Cameron, 2022; Hajkowicz et al., 2021). We ran an unsupervised 
machine learning clustering algorithm2


 to identify locations in Australia where AI companies are geographically clustered. 
We found 8 clusters in Sydney, Melbourne, Brisbane, Perth and Adelaide containing 296 companies (54% of the total 
companies identified). The AI companies are located within proximity to each other, indicating tight clustering, with an 
average nearest neighbour distance of 80 m (range = 55 m to 158 m).

CITY CENTRE

LOW

2 The clusters are centred on these locations and were identified by an unsupervised machine learning algorithm called DBSCAN – density-based spatial 
clustering of applications with noise. We used DBSCAN input parameters of EPS at 400 metres and a minimum sample of 6.



Regional capability strengths

The graphs below show the top five capability categories 
for Australian States and Territories.3


 Except for the 
Australian Capital Territory (ACT), the leading capability 
categories across states were data services and 
finished solutions.

In the ACT, 75% of AI companies had capabilities in skills 
and training, which was greater than the national average 
of 39%. The share of the companies contributing to AI 
capability in consulting was 58% – also above the national 
average. This is likely related to the service focus of ACT and 
a higher concentration of AI companies providing training 
and consulting for government and industry organisations.

South Australia had a higher than national-average 
concentration of research and innovation capabilities. 
This is likely related to the concentration of world-renown 
AI hub of research institutions in Adelaide, including 
Defence Science and Technology Group, the Australian 
Institute for Machine Learning, MITbigdata Living Lab 
by Massachusetts Institute of Technology, the Australian 
Research Centre for Interactive and Virtual Environments, 
and the Australian Cyber Collaboration Centre 
(Bratanova et al., 2022).

Product and service offerings by State and Territory3

Share of product and service 
offering group (%)

National average

3 The Northern Territory and Tasmania are not included in this analysis due to the limited number of companies.



Capability strengths of eight AI clusters

The top five product and service offering categories for 
the eight identified AI company clusters are demonstrated 
below. Data services and finished solutions are among 
the top three categories across all eight clusters. 
Consulting is relatively higher in Perth, Adelaide and 
Brisbane central business district clusters, while the 
Eveleigh cluster in Sydney has over half of member 
companies offering products and services in research 
and innovation.

Product and service offerings by cluster

Share of product and service offering group (%)

National average



Company formation 
and maturity

Australia is home to 544 AI companies, which is comparable 
to other global leading AI countries. For example, having 
a larger population and economy, Canada recently 
reported 670 AI companies (University of Toronto, 2020). 
Until the last decade, the number of AI companies founded 
annually in Australia was growing at a slow pace, with an 
average of 5.5 companies added to the ecosystem annually 
between 1986 and 2013. This growth has since intensified, 
with 74 companies founded in 2017, 61 in 2018 and 57 in 
2019. A similar pattern has been observed internationally, 
including the UK and Singapore.

The cumulative number of Australia AI companies has intensified 
in the last decade

Note: Foundation year data was only available for 531 of the total 
544 Australian AI companies identified in this analysis.

The United Kingdom and Singapore have shown similar 
temporal growth patterns to Australia

The Australian AI business ecosystem is young, with 76% 
of companies (396 companies) founded in the last decade. 
The emerging nature of Australian AI companies is evident 
when compared to the broader Australian IT sector: the 
median company age is 6 years for AI companies and 
13 years for IT companies. The Australian IT sector has 
been growing steadily, with growth intensifying in the 
last decade. Growth in AI companies is skewed more 
sharply towards the last 5 years, with 39.2% of Australian AI 
companies founded in 2018–2022, compared with 14.8% of 
Australian IT companies (Crunchbase, 2023a).

Share of Australian IT and AI companies by founding year

Note: The total count of Australian companies with data on the foundation year was 521 for AI companies and 6,772 for IT companies (1993–2023).



Publicly listed Australian 
AI companies

Our analysis revealed 13 Australian publicly listed AI 
companies trading on the Australian Stock Exchange and 
internationally (UK, USA). The total market capitalisation of 
these companies is approaching $73 billion. All companies 
operate in the IT sector with a specific industry focus. 
Five out of 13 companies develop solutions for the 
healthcare industry, two operate in the fields of energy and 
the environment and one in transportation. While these 
companies operate across multiple Australian offices 
and internationally, their headquarters are in Victoria 
(6 companies), New South Wales (4), Western Australia (2) 
and the ACT (1).

A large majority of Australian AI companies are 
headquartered in New South Wales, with almost every 
second AI company founded in 2013–22 located there. 
These companies are proportionally younger companies 
too. Victoria hosts a larger proportion of more mature 
AI companies, with 40% of Australian AI companies 
founded in 1993–2012 and operating today headquartered 
in Victoria.

New South Wales has a larger share of younger AI companies, with Victoria home to more mature AI companies

Most Australian AI companies have 50 employees or less

Note: The total count of Australian AI companies with data on the number 
of employees was 454 (out of 544 companies included in this analysis).



Jobs and skills demand

The demand for AI-related skills is growing in Australia 
and internationally. In 2022, 2.1% of all job postings in the 
USA were AI-related (up from 1.8% in 2021). Australia is 
among the global leaders in terms of AI job postings, 
with 1.2% of all job postings in 2022 being AI-related. 
Demand for AI jobs has been going faster in Australia 
relative to international comparisons, with the share of 
AI-related job postings increasing by more than 7 times 
between 2014 and 2022.

Share of AI job postings as a percentage of all job postings (%), 2022

Data source: Stanford AI Index report 2023 (Maslej et al., 2023).

Growth rate in AI job postings relative as percentage of all job postings

Data source: Stanford AI Index report 2023 (Maslej et al., 2023).





Investment trends

AI investment has seen substantial growth over the past 
decade globally and in Australia. In 2022, private investment 
in AI globally was $132.3 billion (US$91.9 billion) – 18 times 
higher than it was in 2013 (Maslej et al., 2023). This growth 
dropped off for the first time in 2022, with global private 
investment in AI, declining by 26.7% from 2022 to 2021 
(Maslej et al., 2023). Conversely, private investment in AI 
in Australia spiked in 2021 and remained high in 2022. 
Australia was ranked in 11th place for global ranking 
by private investment in AI in 2022, which reflected an 
improvement from 14th place over the period 2013–21.

Australian investment in AI

Data source: Stanford AI Index report 2023 (Maslej et al., 2023), Global AI 
Vibracy Tool (Stanford HCAI, 2023).

Venture capital (VC) investors 
in Australian AI companies

According to funding deals recorded on Crunchbase, the 
top five investors in Australian AI companies are as follows:

Flying Fox Ventures is a Victoria-based venture capital 
investment firm founded in 2021. The company invests in 
early-stage companies in technology, business-to‑business 
and business-to-consumer sectors in Australia and 
New Zealand.

Artesian VC is a global company with offices in Sydney, 
Melbourne and Adelaide. The company was established 
in 2004 and specialises in public and private debt, VC and 
impact investment strategies.

Blackbird is a VC firm that incubates and invests in 
early‑stage software technology companies.

Main Sequence is a Sydney-based VC firm co-founded by 
CSIRO and the Australian Government in 2017. The firm 
focuses on investment in deep tech and frontier technology 
companies and leveraging world-leading research from 
Australian institutions.

Carthona Capital is a VC investment firm that is based in 
Sydney and invests globally. The firm was founded in 2014 
and focuses on pre-seed, seed and Series A companies.



 2 
Research and 
development

Information about research publishing by Australia’s 
universities and research institutes working on AI 
and information about AI patent application trends.



Methods and source data

For our analysis of research publishing, intellectual property patents and Australian 
research institutes, we draw upon data from The Lens, Scopus by Elsevier and AI 
search phrases from the OECD. We use these datasets to identify AI-related scholarly 
publishing (peer-reviewed books, book chapters, journal papers and conference 
papers) using a bibliometric analysis approach. We apply a similar bibliometric 
analysis as approach used in Hajkowicz et al. (2023b).

OECD

AI search phrases from OECD 
expert groups (225 unique 
phrases/technologies).

Patents 
applications 
in Australia

Total for 2000–23: 
5,281 about AI 
by jurisdiction 
(893,759 in total), 
representing 
0.6% of patents

The Lens

Comprehensive global database on scholarly 
publishing and patents. We extracted data over 
a 24-year period from 2000–23 using the API.

Scopus

All Science Journal Classification Codes (ASJC)

• 4 First level categories
• 31 Second level categories
• 333 Third level categories.


Select patents
and scholarly
publications
from world
and Australia

Scholarly 
publications 
in Australia

Total for 2000–23: 
92,940 about AI 
(1,679,922 in total), 
representing 5.5% 
of publications



Revealed technology advantage metric and statistical test

We adopt the revealed technology advantage (RTA) 
metric to assess the level of specialisation Australia has in 
application domains of AI. The RTA is a metric proposed by 
the OECD (2023) to measure technological specialisation 
and assess the comparative advantages a country or 
a jurisdiction might have in a particular technology. 
Countries/jurisdictions will typically seek to invest in 
technological capabilities where they have a comparative 
advantage (i.e., those denoted by a high RTA).

An illustration of how the RTA is calculated is given below 
for a hypothetical example where a country has 120 AI 
papers in research field X and 1,000 papers in research field 
X in total. This compares to the world which has 30,000 AI 
papers in research field X and 1,000,000 papers in research 
field X in total. This would yield a high RTA score of 4. 
The calculations are as follows:

While there is no clear cut-off about what represents a 
significant RTA, a value of 4 or greater would typically 
indicate an area of highly specialised technological 
capability. In this hypothetical case we would probably 
infer that the country has specialised AI capability in 
research field X. RTAs are similar to location quotients 
in economic geography where values above 1.2 to 1.5 are 
often considered evidence of significant specialisation 
(Crawley et al. 2013).

In our analysis, we apply a statistical technique developed 
by Crawley et al. (2013) to identify ranges above and below 
the estimated RTA score. Originally developed for use in the 
calculation of location quotients in economic geography, 
this statistical technique handles cases where a small 
number of publications might yield a high score but also 
carries high uncertainty and is therefore not significant.

RTA

=

Number of AI papers in field_x in country

Number of papers in field_x in country

Number of AI papers in field_x in world

Number of papers in field_x in world

RTA

=

120

1000

30,000

1,000,000

RTA

=

12%

3%

RTA

=

4.0



Knowledge creation 
(research publishing trends)

Since 2000, Australian researchers have published 
92,940 peer-reviewed journal articles, conference papers, 
books and book chapters that refer to AI technologies in 
the title, abstract or keywords. This represents 5.5% of our 
total publishing. The intensity of AI publishing has increased 
over time, with AI publishing accounting for 9.6% of all 
publishing today compared with 2.5% in 2000.

Compared to other countries Australian researchers have 
higher AI adoption rates, with 9.6% of Australian research 
publications using or developing AI technology versus a 
global average of 7.2%. Since 2000, Australia created 2.2% 
of AI-related global publications, compared with 1.5% of 
global research publications on all topics, demonstrating 
Australia’s above-average contribution to global AI research. 
While Australia’s publications grew in absolute terms from 
2021 to 2022, Australia’s share of global publications declined 
due to a rapid increase in global publishing driven by other 
countries, such as China.

Share of all peer-reviewed publishing on the topic of AI

Australia’s share of global peer-reviewed publishing

The volume of AI research publications with an Australian author 
has been growing, except for the pandemic when there was a 
downturn in publishing in most fields

Note: According to The Lens, overall publishing in all fields saw a slowdown 
during 2021–22 associated with the COVID-19 pandemic. Data for 2023 is an 
incomplete year with data up until June 2023.

AI publishing intensity has increased across all fields of research 
(first-level categories) in Australia, with the strongest growth in 
physical sciences

Note: The last data point is for 2023 and is based on data only from January 
to June 2023.

AI publishing 
intensity

Number of peer-reviewed 
AI publications in field

Total number of all peer-reviewed 
publications in field

....
NUMBER OF AUSTRALIAN AI PUBLICATIONS  

The footprint of AI in Australia across more granular subject groups

Over the past two decades, AI has established a strong 
position in practically all research fields in Australia, 
suggesting that Australian researchers are making use of 
AI to further innovation and problem-solving in their own 
field. Across the second-level ASJC subject fields, the areas 
that had the highest level of AI research publishing in 2023 
were computer science, mathematics, decision sciences, 
environmental science and engineering.

Whilst these are research fields, they’re likely to have 
some connection to cutting-edge technology innovation 
within associated Australian industries. For example, 
the “health professions” field is likely to capture much 
AI activity within the healthcare sector in Australia.

Percent share of Australian research publishing by second-level ASJC field

ALL SCIENCE JOURNAL CLASSIFICATION 
(LEVEL 2)

2000

2005

2010

2012

2014

2016

2018

2019

2020

2021

2022

2023*

Agricultural and Biological Sciences

0.9%

1.3%

1.7%

2.0%

2.5%

3.0%

3.6%

3.9%

4.3%

4.4%

6.6%

6.0%

Arts and Humanities

2.2%

1.6%

1.7%

2.1%

1.4%

1.8%

1.6%

2.5%

2.5%

2.6%

3.9%

2.8%

Biochemistry Genetics and Molecular Biology

0.7%

1.3%

2.2%

2.7%

3.6%

4.0%

4.4%

4.6%

5.4%

5.9%

7.5%

7.7%

Business Management and Accounting

2.9%

1.1%

2.9%

2.9%

2.5%

4.5%

4.6%

4.9%

5.4%

7.0%

6.4%

7.4%

Chemical Engineering

2.0%

1.9%

1.1%

1.5%

2.3%

2.6%

1.6%

2.3%

2.2%

2.7%

5.1%

5.5%

Chemistry

0.8%

1.4%

1.0%

1.4%

1.5%

1.7%

1.9%

2.3%

2.6%

3.6%

4.8%

4.6%

Computer Science

20.2%

22.7%

21.4%

24.1%

24.5%

27.0%

31.1%

32.7%

35.0%

39.5%

39.2%

40.3%

Decision Sciences

7.4%

7.4%

11.0%

11.8%

8.3%

12.7%

13.6%

12.6%

16.6%

18.7%

20.1%

21.3%

Dentistry

1.3%

3.2%

3.0%

2.0%

5.7%

2.3%

3.6%

3.4%

3.4%

8.7%

6.7%

4.0%

Earth and Planetary Sciences

0.9%

1.8%

1.7%

2.0%

2.9%

3.2%

3.9%

4.4%

4.8%

5.8%

8.0%

8.1%

Economics Econometrics and Finance

1.3%

2.1%

1.1%

1.0%

1.0%

1.2%

1.6%

2.5%

2.2%

3.1%

5.2%

4.3%

Energy

0.5%

3.1%

2.6%

3.1%

3.7%

4.5%

3.1%

4.5%

5.2%

5.9%

7.0%

9.5%

Engineering

5.1%

4.8%

5.3%

6.9%

7.0%

7.8%

8.5%

10.8%

11.8%

14.0%

15.0%

19.8%

Environmental Science

1.5%

2.2%

2.0%

2.2%

2.8%

3.1%

4.1%

4.4%

4.5%

5.1%

6.3%

6.3%

Health Professions

2.3%

0.7%

1.8%

2.3%

3.5%

3.6%

4.1%

5.5%

5.5%

7.5%

9.4%

10.7%

Immunology and Microbiology

0.5%

1.1%

2.0%

2.4%

2.9%

3.0%

2.7%

4.1%

3.9%

4.5%

5.4%

3.3%

Materials Science

0.8%

0.7%

0.8%

1.1%

1.1%

1.6%

1.8%

3.8%

3.9%

4.2%

4.7%

5.3%

Mathematics

8.7%

13.5%

12.9%

13.8%

12.8%

15.4%

16.2%

17.7%

18.9%

20.7%

24.1%

25.1%

Medicine

1.4%

1.8%

2.5%

3.1%

3.5%

4.1%

4.7%

5.2%

5.7%

6.5%

8.0%

8.1%

Neuroscience

1.6%

2.8%

3.5%

4.9%

5.4%

6.0%

6.4%

6.5%

7.6%

8.2%

10.1%

10.4%

Nursing

1.6%

0.9%

2.3%

3.2%

2.8%

3.7%

4.6%

4.5%

4.7%

5.0%

6.3%

6.2%

Pharmacology Toxicology and Pharmaceutics

0.7%

1.1%

1.5%

1.6%

2.3%

1.5%

3.9%

3.5%

4.3%

4.7%

5.8%

4.4%

Physics and Astronomy

1.4%

1.5%

1.4%

1.4%

1.9%

2.3%

3.6%

4.8%

5.4%

7.6%

8.1%

10.1%

Psychology

2.8%

3.6%

3.6%

5.6%

4.1%

5.4%

4.9%

5.1%

5.3%

4.9%

6.1%

6.2%

Social Sciences

1.1%

1.5%

1.5%

1.8%

1.7%

2.3%

2.4%

3.2%

3.5%

3.6%

5.4%

5.2%

Veterinary

0.4%

1.0%

1.8%

3.7%

3.4%

4.8%

3.3%

4.2%

4.0%

5.1%

7.5%

3.9%





*Includes data for January to June 2023.



Application domain 
specialisations

Of the 333 third-level ASJC research fields where AI is 
being applied, we identified 31 fields that reflect areas of 
significant specialisation for Australia and likely areas of 
comparative advantage. Areas of significant specialisation 
were identified as research fields with Australian research 
publishing RTA of 1.3 or greater and a lower bound of the 
uncertainty range of 1.1 at the 95% confidence interval. 
We note the cut-offs are assumed; in reality, the level of 
specialisation is a continuously graded scale from high to 
low. We use the cut-offs to create a shortlist of areas of 
higher specialisation. There were 31 application domains 
with an RTA score that met the criteria for significant 
specialisation (above 1.3) and had statistical significance 
at the 95% confidence interval.

Specialisation in AI application domains in 2022



For the 31 application domains with significant 
specialisation for Australia, we calculated the share of 
research in this field globally that is driven by an Australian 
author. Australian authors account for approximately 
26% of global AI publications in optometry, 8.5% in 
livestock production and 6.8% in the decision sciences, 
which is greater than the average Australian share 
of global research publishing across all fields (1.3%). 
Optometry, livestock production and decision sciences 
therefore reflect AI application domains where Australian 
research is highly represented.

Australia’s share of global AI publishing in application domains in 2022



Artificial intelligence 
technology specialisations

We find numerous AI technologies where Australia has specialised capability on the global stage. Here we show 
AI technologies with an RTA > 1.5 and lower bound above 1.0 at 95% Confidence Interval. Data is for the year 2022.

AI TECHNOLOGY

DESCRIPTION (TEXT IN THIS FIELD WAS INITIALLY GENERATED BY 
CHATGPT 4 AND THEN CHECKED AND EDITED BY THE LEAD AUTHOR)

RTA

AUSTRALIAN 
SHARE OF GLOBAL 
PUBLISHING (%)

Self-organising map

A type of artificial neural network trained using unsupervised machine 
learning to produce a low-dimensional representation of the input space, 
typically a 2D grid.

7.63

14.0

Ant colony optimisation

A type of optimisation inspired by the behaviour of ant colonies. 
It is mainly used to find optimal paths through maps or charts.

5.27

9.7

Gradient tree boosting

A machine learning technique that produces a prediction model in the form 
of an ensemble of weak prediction models, typically decision trees.

5.19

9.5

Multi-objective 
optimisation

An area of multiple criteria decision analysis concerned with mathematical 
optimization problems involving more than one objective function to be 
optimized simultaneously.

3.03

5.6

Evolutionary 
computation

A family of algorithms for optimization inspired by biological evolution, such 
as genetic algorithms, evolutionary strategies, genetic programming, etc.

2.58

4.7

Variational inference

A method in machine learning that uses optimization techniques to estimate 
the true probability distribution of data.

2.50

4.6

Social robot

A type of autonomous robot that interacts and communicates with humans or 
other autonomous physical agents by following social behaviours and rules.

1.96

3.6

Recommender system

This is a type of machine learning that uses information provided by a user to 
predict and/or prioritise the products/items they’re seeking.

1.95

3.6

Link prediction

A problem in network science and social network analysis to predict the 
existence of a link between two nodes, given a snapshot of a network.

1.89

3.5

Cognitive modelling

A method used in AI to simulate human problem-solving and mental task 
processes in a computerized model.

1.86

3.4

Evolutionary algorithm

A subset of evolutionary computation, a generic population-based 
metaheuristic optimization algorithm.

1.68

3.1

Human-robot 
interaction

A field of study dedicated to understanding, designing and evaluating robotic 
systems for use by or with humans.

1.59

2.9

Meta learning

These are models that learn how to learn. The learning algorithm is adjusted 
and improved in each iteration of problem solving. It helps identify which 
algorithms perform best for a given problem.

1.58

2.9

Decision model

This captures a wide range of AI-based tools and technologies which 
are designed to help a decision model choose from a set of competing 
alternatives.

1.54

2.8

Random field

Random fields are used to infer the joint distribution of a set of variables 
based on their interactions and dependencies. They’re used in image 
processing, language processing and computational biology.

1.51

2.8







Artificial intelligence technology specialisations – error margins

This graph shows the data on AI technology specialisation in Australia from the previous table but with 
error margins around the RTA scores at the 95% confidence interval.

Error margins at 95% confidence interval for AI technology specialisations in 2022

Australia has an above-average share of global AI publishing in AI technologies where it has significant specialisation, 2022

.......................
Random 
eldDecision modelMeta learningHuman-robot interactionEvolutionary algorithmCognitive modellingLink predictionRecommender systemSocial robotVariational inferenceEvolutionary computationMulti-objective optimisationGradient tree boostingAnt colony optimisationSelf-organising mapAUSTRALIAN SHARE OF GLOBAL AI PUBLISHING ”%–
Australia accounts for .%
of global research publishingon all topics in all 
elds

Universities and 
research institutes

Our datasets identify 167 research institutes in Australia with a unique global Research 
Organisation Registry (ROR) code. Twenty-five of these institutes had published over 
100 peer‑reviewed journal papers, books, book chapters or conference proceedings on 
AI in 2021. Moreover, six institutes referred to AI technologies in over 10% of publications, 
including the University of Technology Sydney, which had the highest AI publishing 
intensity among Australian universities and research institutes.

Intensity of AI publishing by research institute as indexed by the share of total publishing on AI, 2021

Note: This analysis was based on 2021 data from The Lens due to declines observed across many institutes in 2022, which was likely associated with the 
COVID-19 pandemic. We assume that 2021 is a more representative year of AI publishing activity than 2022.



When listed in descending order of volume of AI publishing 
the ranking of institutes (with over 50 AI papers in 2021) 
changes. The University of Sydney, University of New South 
Wales, Monash University, University of Melbourne and 
the University of Technology Sydney all contributed over 
1000 peer reviewed AI publications in the year 2021.

University research centres, 
industry and community groups 
in the Australian AI ecosystem

Australian universities have research departments and 
groups dedicated to AI research, many of which were 
founded in partnerships with the Australian Government, 
regional governments and industry organisations. 
Research, innovation and commercialisation activities by 
the Australian universities contribute to the development 
and expansion of the AI ecosystem. The table below 
provides examples of the university research groups and 
departments that specialise in AI.

Australia is home to multiple industry and community 
groups that actively contribute to the AI ecosystem, 
especially in relation to skills and training, research and 
innovation. The table below provides some examples.

Total number of AI publications in 2021 by research institute (institutes with over 100 AI publications)

Note: This analysis was based on 2021 data from The Lens due to declines observed across many institutes in 2022, which was likely associated with the 
COVID-19 pandemic. We assume that 2021 is a more representative year of AI publishing activity than 2022.



University research groups and departments that specialise in AI

UNIVERSITY

DEPARTMENT OR RESEARCH GROUP

University of Western Australia

UWA Data Institute

Monash University

Department of Data Science and Artificial Intelligence

Australian National University

Intelligent Systems

University of Adelaide

Australian Institute for Machine Learning

University of Melbourne

Artificial Intelligence Assurance Lab and AI and Autonomy Lab
(School of Computing and Information Systems)

University of New South Wales

UNSW AI Institute

University of Queensland

UQ AI Collaboratory

The University of Sydney

Sydney Artificial Intelligence Centre

Sydney Institute for Robotics and Intelligent Systems Australian Centre for Field Robotics

Macquarie University

Centre for Applied Artificial Intelligence

Deakin University

Centre for AI and Future of Business

La Trobe University, CISCO

AI & Machine Learning at La Trobe University

RMIT

Centre for Industrial AI Research & Innovation

University of Technology Sydney

Australian Artificial Intelligence Institute

Torrens University Australia

Centre for Artificial Intelligence Research and Optimisation





Industry and community groups that actively contribute to the AI ecosystem

NAME OF ORGANISATION OR INITIATIVE

HEAD ORGANISATION OR FOUNDING ORGANISATIONS

ESTABLISHED

National Artificial Intelligence Centre

CSIRO, Google, Committee for Economic Development of Australia

2021

AcademicID

CSIRO

2022

Artificial intelligence Ethics Committee

Australian Computer Society

NA

ARC Centre of Excellence for Automated 
Decision‑Making and Society

Australian Research Council, Australian Government

2020

Advanced Robotics for Manufacturing Hub

Queensland University of Technology, Urban Art Projects

2019

Australian Urban Research Infrastructure Network

University of Melbourne

2022

Australian Council of Learned Academies

Five Learned Academies

2010

Artificial Intelligence Laboratory

University of Adelaide

2017

Advanced Analytics and AI Platform

Intersect Australia

NA

Gradient Institute

IAG, The University of Sydney, CSIRO

2018

Queensland AI Hub

Queensland Government, University of Queensland, 
Queensland University of Technology

2020

Data Science and Ai Association of Australia

NA

2018







Product innovation and 
patent applications

Intellectual property patent applications can be used as 
an indicator that the applicant – an individual, company 
or organisation – believes they have discovered a novel 
technology with commercial value. Patent applications 
are costly and time-consuming, so they have to be worth 
the effort. Analysing AI patents provides insights into the 
creation of cutting-edge commercially valuable inventions 
in Australia.

However, we note that patents only capture part of what’s 
happening. Research has shown that 36% of product 
innovations result in patent applications (Arundel and 
Kabla, 1998). So, whilst patent analyses provide some 
insight into AI and product innovation, they likely 
do not capture the bulk of activity, nor is there an 
alternative comprehensive and consistent measure of 
product innovation.

With these caveats in mind, patents can still provide 
insights into AI product innovation trends. In this section, 
we analyse patent data relating to AI for Australia. 
The headline graph on the number of AI patents by 
Australian jurisdiction, applicant and inventor shows the 
bulk of Australian AI patents had an overseas applicant or 
inventor. All three trends show growth patterns, with more 
rapid growth since 2016.

The share of global AI patent applicants and inventors 
in Australia has fallen over time. This trend is likely to be 
largely driven by increases in global patenting activity, 
rather than a decline in patenting activity in Australia. 
However, the number of global AI patent applications in the 
Australian jurisdiction has grown sharply since 2010, which 
is likely to be associated with the increasing presence of 
overseas technology companies based in Australia.

AI patent applications in Australia

There has been a sharp increase in the share of global AI patent 
applications submitted in the Australian jurisdiction

Note: The last data point is for 2023 and is based on data only from January 
to June 2023.



There were 379 AI patent applications in the Australian 
jurisdiction in 2022 (the last full year of data in The Lens), 
compared with 73 submitted by Australian applicants and 
119 by Australian inventors. The companies filing for the 
largest number of AI patent applications in Australia have 
their global headquarters in other countries. The top 10 
companies submitting AI patent applications in Australia 
account for 32% of all AI patent applications filed in the 
Australian jurisdiction from 2000–22.

Examining AI patent applications by inventors (as opposed 
to by applicant or jurisdiction) provides insights into the AI 
product innovations being created by Australian inventors. 
The data show a sharp rise in relative and absolute terms 
since around 2015. In 2023, 3.6% of patents invented by 
an Australian resident were about AI and 203 AI patents 
were invented by Australians in 2022 (2.7% patents). 
These temporal patterns show that AI technology is 
playing an increasingly important role in Australian 
product innovation.

Top 10 companies that submitted AI patent applications in the Australian jurisdiction during 2000–22, all of which have international 
headquarters

Australian product innovation is increasingly incorporating 
AI technology

Note: The last data point is for 2023 and is based on data only from January to June 2023.

Overall volume of AI product innovation has risen sharply 
since 2015


NUMBER OF AI PATENT APPLICATIONS IN AUSTRALIASUBMITTED BY AN AUSTRALIAN INVENTOR 

We analysed which of the 225 AI phrases identified by the OECD as relating to AI technologies 
occurred most frequently in AI patent applications in Australia. Terms such as “robot”, 
“image processing”, “machine learning”, “neural network” and “learning model” were most 
frequently used in patent descriptions in patent applications with an Australian inventor. 
The chart below shows the top 20 AI technology phrases that featured in patent applications 
with an Australian inventor occurring during 2000 to 2023.



The share of global AI patents with an Australian inventor 
has risen strongly in the last year, increasing from 0.11% in 
2022 to 0.24% in 2023. This last data point captures only 
January to June 2023; whether it’s an aberration or longer 
term trend is unclear. This share of global AI patents is on 
par with the share of global patents covering all topics 
that have an Australian inventor. Australia’s share of global 
patents (on all topics) has been on the decline due to 
the rapid growth of patent production, especially from 
emerging economies such as China.

Australia’s contribution to global AI patent applications 
remains well below our contribution to AI research 
publishing. From January to June 2023, Australia contributed 
1.6% of the world’s peer-reviewed AI research publishing 
but only 0.24% of global AI patent applications (with an 
Australian inventor). This discrepancy suggests that we are 
not converting AI research into AI product innovations and 
commercialisation opportunities at the same rate as other 
countries. We create much more knowledge about AI for 
the world than we do commercial AI products.

Share of global patents invented by Australian residents

Note: The last data point is for 2023 and is based on data only from January to June 2023.

Australia makes up a larger share of global research outputs 
than patent applications

0.24% of global AI 
patent applications 
had an Australian 
inventor during 
January to June 2023



 3 
Industry stakeholder 
and expert perspectives

Qualitative analysis of interviews conducted by the 
research team about opportunities and challenges 
for Australia’s AI ecosystem.



Insights from our interviews

We conducted consultations with 28 stakeholders across 
the AI ecosystem in Australia, including representatives 
from companies developing and adopting AI technologies 
and supporting services, government agencies and 
academic institutions. These engagements sought 
to better understand current and emerging areas of 
competitive advantage across the Australian AI ecosystem, 
the unique value proposition of Australian-developed AI 
and opportunities to grow and support the ecosystem in 
the future. This section details the key themes emerging 
from these consultations and their implications for future 
directions across the AI ecosystem in Australia.

The hype versus the reality 
around AI

Interest in AI technologies has been accelerated by the 
release of tools like OpenAI’s ChatGPT, Google’s Bard and 
Microsoft’s Bing and there is a sense that we are still in 
the beginning stages of the current AI hype cycle. These 
developments have made it easier for business leaders to 
visualise the potential benefits that AI technology could 
provide for their organisations and Australian AI service 
providers are routinely using ChatGPT as a useful reference 
point in discussing AI opportunities with clients.

Despite this interest, there is a reasonable degree of 
apprehension across government and businesses about 
adopting and experimenting with AI technologies, with 
some organisations more risk-averse than others. Moreover, 
even if a business is open to implementing AI technologies, 
many are held back due to fragmented data systems and/or 
the investment needed to resolve these data issues. Further 
work is needed to educate decision-makers about the types 
of problems that AI can address or are best suited to and 
the level of investment and development required to reach 
an optimal solution.

The perceived value of local AI 
service providers

When it comes to deciding whether to outsource AI 
capabilities to an Australian or international provider, 
there are several considerations. A key factor is access 
to capabilities, with many companies opting for an 
international AI service provider if the appropriate 
capabilities do not exist or cannot be identified locally. 
The NAIC has established the Australian AI Ecosystem 
Discoverability Platform, a directory of Australian 
capabilities that support the adoption of AI technologies 
and a useful starting point in connecting local AI service 
providers with businesses (CSIRO, 2023). There could be 
future opportunities to expand the platform’s coverage of 
the Australian AI ecosystem and awareness of the platform.

There are natural proximity advantages to doing business 
with Australian AI service providers. These include 
operating in the same time zone, having access to face-
to-face support and an appreciation for the local industry 
context. There is also an innate trust associated with 
Australian-developed AI technologies (see An Opportunity 
for safe and responsible AI). On the other hand, there 
is a tendency to assume that AI models that have been 
developed by larger international companies are more 
reliable or scalable by virtue of being developed in bigger 
markets or with greater resources. This perception, along 
with existing business relationships with international 
suppliers, can sway Australian companies to do business 
with an international AI service provider.

Deciding on what AI to build, 
buy or borrow

Will Australia be a leading creator of AI technologies 
or adopt technologies developed elsewhere? There is a 
general perception across the AI ecosystem that Australian 
businesses will likely be a consumer of internationally 
developed AI technologies. Despite being the 13th most 
advanced economy, Australia is currently ranked as the 
82nd most complex economy (Observatory of Economic 
Complexity, 2023) and has low rates of adoption of AI and 
data analytics relative to other OECD countries (Productivity 
Commission, 2022). To be globally competitive, Australia 
needs to set the conditions that support AI adoption as well 
as create unique AI capabilities.

Australia has a strong network of universities and 
research institutes working on AI technologies, but 
there is a persisting challenge in commercialising and 
scaling innovations. There are likely to be cases where 
it is advantageous to outsource or adapt existing AI 
technologies when there are cost and capabilities barriers 
to developing from scratch. But there could be areas where 
there is an imperative to develop sovereign foundational AI 
models and technologies (e.g., applications based on data 
resources that cannot be used to train international models 
due to privacy restrictions). Some stakeholders were 
supportive of using open-source models as a starting point 
for building these foundational AI capabilities.



Australia’s modest culture and low appetite for risk 
were also highlighted as factors that are holding the AI 
ecosystem back from developing novel AI technologies. 
This was exemplified in expatriates who had returned 
to Australia after spending a substantial period working 
in international technology clusters like Silicon Valley. 
Showcasing AI success stories of innovative AI technologies 
and applications across the Australian AI ecosystem, such as 
those featured in this report, could help to create a cultural 
shift around opportunities for the Australian AI ecosystem 
to drive future technology developments.

Prioritising being an AI specialist 
over an AI generalist

Australian businesses felt the current national AI approach 
is too broad and aspirations to position Australia as a 
global leader in AI technologies could be misplaced. This 
sentiment is based on the acknowledgement that we do 
not have the population, market or economy size and 
level of investment in AI needed to compete with large 
multinational technology companies. Instead, a more 
desirable approach could be to focus on building and 
growing our AI capabilities and reputation in specialised 
areas where Australia has a strategic and/or comparative 
advantage. Deciding on areas of AI specialisation could take 
different forms.

For example, there could be opportunities to leverage 
existing industry strengths in heavy industries, such as 
mining and agriculture, where there are clear benefits for 
the role of automation, computer vision and advanced 
decision support tools. There could also be opportunities 
to leverage Australia’s rich renewable energy resources to 
sustainably power AI technologies. Australia’s large, open 
spaces suitable for testing novel autonomous technologies 
could also be attractive for international companies looking 
to develop and test in a smaller market. Finally, in line with 
the Australian Government’s ambition for Australia to be 
a global leader in trusted, secure and responsible AI (DISR, 
2023a), stakeholders acknowledged the unique advantage 
of Australia in developing trusted AI technologies (see An 
Opportunity for safe and responsible AI).

An opportunity for safe and 
responsible AI

Ensuring safe and responsible AI practices is important for 
Australian businesses to mitigate potential risks associated 
with AI, including legal and reputational risks (Reid et al., 
2023). It is acknowledged across the Australian AI ecosystem 
that there is a significant opportunity for Australia to be 
a global leader in the development and use of safe and 
responsible AI technologies. This ambition aligns with 
the Australian Government’s latest consultation on safe 
and responsible AI in Australia which aims to identify 
the optimal regulatory and policy approach to ensure AI 
developed and used in Australia is done so in a safe and 
responsible manner (DISR, 2023b).

Australia is not alone in its desire to be a global leader 
in safe and responsible AI. Other countries, including 
Singapore, the United States, the United Kingdom and the 
European Union, are among a suite of jurisdictions that 
are exploring both voluntary and regulatory approaches 
safeguarding future AI developments and applications 
(DISR, 2023b). Stakeholders across the Australian AI 
ecosystem emphasised the need for a balanced, risk-
based approach to regulating AI to ensure such policies 
do not unnecessarily slow or disrupt future AI uptake and 
developments.

The ‘Australian brand’ is a trusted asset, which puts 
the nation in a strong position to drive future safe and 
responsible AI developments. There is a current disconnect, 
however, between best practice approaches to responsible 
AI and actual business practices: the 2023 Responsible AI 
Index found that

82% of Australian organisations believed they were 
practising AI responsibly but only 24% had measures in 
place to ensure this was the case (Fifth Quadrant, 2022). 
To support Australian businesses in implementing AI 
safely and responsibly, the NAIC and the Gradient Institute 
have released a guide on implementing the Australian 
Government’s eight AI ethics principles (Reid et al., 2023).



Finding our strength through 
our diversity

With every 3 in 10 people living in Australia born overseas, 
Australia is one of the most culturally and linguistically 
diverse countries in the world (AIHW, 2023). This provides 
a unique comparative advantage when it comes to 
developing novel AI technologies: AI models that are 
trained on representative Australian data are likely to be 
less vulnerable to population biases and be fit for purpose 
as an exportable AI product or service in other countries. 
This data diversity could be particularly powerful when 
developing AI health applications or designing technologies 
to mitigate biases against marginalised populations.

Conversely, a lack of diversity in training data used 
in developing international AI technologies can be a 
challenge for Australian businesses looking to apply these 
technologies locally given they might not perform well 
in an Australian context. There is an opportunity for the 
Australian AI ecosystem to leverage this diversity in the 
creation of novel AI technologies, as well as position itself 
as the testbed for international companies looking to 
improve the quality and representativeness of the training 
data their models are built upon.

Strengthening linkages across the 
AI ecosystem

Across the AI ecosystem in Australia, there is a perceived 
lack of strong leadership and connectivity. In some cases, 
this can lead to duplicated efforts across academia and 
industry, with parallel streams of AI development on 
similar technologies. In other cases, Australian businesses 
can struggle to identify the right AI capabilities they need 
and resort to sourcing these capabilities internationally. 
While these barriers likely reflect the nascent nature of 
Australia’s emerging AI sector, it nonetheless requires 
attention to maximise opportunities for Australia to define 
and strengthen its global reputation in AI and mature as 
an ecosystem.

Some stakeholders pointed to opportunities to leverage 
various national and state-based AI hubs and networks 
across Australia as a mechanism for improving connectivity 
across the ecosystem. Example AI hubs across Australia 
include the NAIC (and its Responsible AI Network), the 
Queensland AI Hub, and the university and research 
institutions listed in this report (see Universities and 
research institutes).

Partner programs offered by large technology companies, 
such as Amazon Web Services Partner Programs and 
Google’s Cloud Partner Advantage Program, were also 
acknowledged as a useful channel for increasing the reach 
of Australian AI service providers, particularly startups and 
smaller providers.

Growing Australia’s AI talent and 
business ecosystem

Talent shortages are a persisting struggle reported by 
Australian businesses, particularly when it comes to 
sourcing specialised technical skills. While it can be 
challenging for Australian businesses to offer globally 
competitive salary packages, the liveability benefits of 
Australia can be attractive for skilled workers looking to 
stay in Australia or relocate here. Beyond these lifestyle 
perks, the stakeholders consulted in this project suggested 
other potential mechanisms for growing Australia’s AI 
talent pipeline and attracting and retaining AI companies 
in Australia.

For instance, countries such as Singapore have introduced 
several top-down measures designed to grow their AI 
sector as part of their national AI strategy, including 
streamlining its process for patenting AI technologies, 
boosting research and development investment into AI 
and fostering an attractive AI startup ecosystem (Goode 
et al., 2023). Australian expatriate technology workers can 
also face challenges returning home to work remotely for 
an international technology company, with stakeholders 
acknowledging the need to reduce the friction in this 
process. Australia could also consider new approaches 
for injecting talent into the AI ecosystem, such as hybrid 
academic-industry positions, which could also help to 
improve connectivity across sectors.



Australia’s AI ecosystem 
case studies

The following case studies provide examples of Australian companies 
at the coal-face of developing AI solutions for industry.

Streamlining the management of critical infrastructure

There are thousands of kilometres of wastewater pipelines 
beneath us that need to be regularly inspected for 
maintenance or repair issues. Traditionally, engineers 
would inspect the condition of these pipes manually using 
CCTV footage, which is extremely challenging and costly 
for utilities, councils and organisations. Sydney-based 
VAPAR uses AI to eliminate the repetitive and manual 
requirements of infrastructure management (VAPAR, 
2023). Its cloud‑based platform uses machine learning to 
automatically assess the condition of pipes and identify 
defects from CCTV footage.

The VAPAR platform improves the accuracy and timeliness 
of pipe condition assessments, generates cost savings 
and strengthens the consistency across assessments. 
VAPAR processes over 200,000 metres of wastewater 
infrastructure across Australia, New Zealand and the United 
Kingdom. This case study highlights Australia’s ability to 
develop AI solutions that solve tangible problems improving 
the safety and efficiency of critical infrastructure.

Image source: VAPAR.Solutions



Helping clinicians detect conditions better and faster

Sydney-based harrison.ai uses AI to address the healthcare 
system’s wicked problem – the need to provide quality, 
equitable and timely care in the face of growing workforce 
shortages (Harrison AI, 2023). Already caring for over one 
million patients a year in Australia, their goal is to raise 
the standard of healthcare for one million patients a day 
by 2025 and make equitable healthcare a reality for all. 
harrison.ai is developing technologies that combine the 
strengths of human intelligence and AI and equip clinicians 
with decision-support tools that can improve their capacity 
and quality of care.

In partnership with I-MED, harrison.ai launched annalise.
ai in 2020, a decision-support solution for chest X-rays. 
The first product, Annalise CXR, provides decision support 
for radiologists to detect conditions quickly and accurately 
from chest X-rays (CXR). This solution outperforms similar 
CXR systems and can detect more clinical findings (124) 
than the next most comprehensive CXR product (75). 
A peer‑reviewed study of Annalise CXR also found a 
significant improvement in radiologist reporting, with the 
AI model matching a radiologist’s decision in 86.5% of cases 
(Jones et al., 2021). franklin.ai, a joint venture between Sonic 
Healthcare and harrison.ai, aims to similarly augment and 
improve pathologists’ decision-making using AI. 

Image source: harrison.ai



Taking a human-first approach to AI in recruitment

Melbourne-based Sapia.ai has developed an AI-enabled 
recruitment tool that aims to make hiring more inclusive, 
efficient and effective (Sapia, 2023). Sapia.ai uses 
proprietary algorithms and data to automate elements 
of the hiring process, reducing time-to-hire for large 
organisations, interrupting bias in hiring and enhancing 
the candidate experience.

One of Sapia.ai’s key tenets is that the end-to-end system, 
including the algorithms, must be fair and explainable. 
Their platform was built to minimise bias in the hiring 
process and excludes the use of video data, data scraped 
from the web, metadata or any other third-party data. 
The algorithms only use data provided by the candidate 
with consent. They have also developed and published a 
FAIR AI for Recruitment (FAIR) framework, which presents 
a set of guidelines for fair, transparent and trustworthy 
applications of AI in recruitment.

The platform has been used to interview more than 
3 million people in 47 countries, including candidates 
from industries such as retail, insurance, financial services, 
healthcare and aviation. Sapia’s national and international 
success provides an exemplar of the value of fair, trusted 
and responsible AI technologies and lays the groundwork 
for future opportunities to position Australia as a leader 
in this domain.

Image source: Sapia.ai



Mining data for new mineral opportunities

Australia’s renowned mining sector is poised to harness the 
potential of technological advancements, with AI playing 
a central role. In 2018 an Australian technology company 
and Solve Geosolutions, an Australian data science 
consultancy, came together to create Datarock (DiUS, 2023). 
They saw an opportunity to leverage AI technology and 
provide the industry with a solution that transforms core 
imagery into actionable data and delivers reliable insights 
(Datarock, 2023; DiUS, 2023).

Traditionally, drill core is logged by manual inspection by 
a geologist or geotechnical engineers, a labour‑intensive 
approach susceptible to inconsistencies and errors. 
Datarock technology – which utilises deep learning 
models and computer vision – enables core imagery to 
be processed via a cloud-hosted platform, automating 
drill core analysis and extracting geoscientific insights 
that are consistent, auditable, and can surpass the quality 
and quantity of traditional human-based observations.

The platform eliminates mundane human effort, 
empowering the user to focus on data interpretation and 
making informed, data-rich decisions that add additional 
value to mining operations. Datarock has been embraced 
by many of the world’s largest mining companies across 
the globe.

Image source: DiUS and Datarock



Connecting remote communities with essential supplies and services

Swoop Aero has developed the world’s first end‑to‑end 
drone logistics platform that is working to bridge 
the tyranny of distance across Australia and beyond 
(Swoop Aero, 2023). Using digital twins, sensor 
technologies, machine learning and computer vision, 
Swoop Aero’s drones are plugged into an integrated 
logistics network to seamlessly deliver products of up to 
5 kg, including medical supplies, with a range of 130 km 
at speeds of 200 km/h. This technology has also been 
used to map and respond to disaster-prone areas, patrol 
coastal environments for incidents and monitor wildlife 
and flora for conservation purposes.

In 2018, Swoop Aero was awarded the world’s first 
commercial contract to provide medical drone logistics 
services in Vanuatu. Healthcare workers previously 
transported vaccinations in iceboxes and would have to 
navigate difficult terrain to deliver these medical supplies. 
Swoop Aero successfully piloted an integrated drone 
network for delivering vaccines and other medical supplies 
to remote locations across Vanuatu and has since operated 
over 70 flights and delivered more than 25 kg of medical 
supplies. Similar successful healthcare projects have been 
conducted in Malawi, the Democratic Republic of Congo 
and Mozambique, providing timely access to medical 
supplies for communities living in remote locations.



Key takeaways

Australia’s AI companies have 
grown rapidly in the last five years

Until the last decade, the number of AI companies 
founded annually in Australia was growing at a slow pace. 
An average of 5.5 companies were added to the ecosystem 
per year between 1986 and 2013. The last decade saw 
intensified growth in the number of newly registered 
Australian AI companies, with a spike of 74 companies 
founded in 2017 followed by 61 and 57 over 2018–19. 
We now have 544 AI companies, which is on par with other 
global AI leaders. For example, with a larger population 
and economy, Canada recently reported 670 AI companies 
(University of Toronto, 2020).

Australia’s AI companies are 
young, vibrant and nimble

Three-quarters of Australia’s AI companies have been 
operating for less than 10 years. This is comparatively 
younger than both economy-wide companies and IT 
sector companies. For example, around 40% of Australian 
AI companies were founded in 2018–22 compared 
to roughly 15% of Australian IT companies. Our AI 
companies are also small, often employing 50 staff or less. 
However, the youth and size of Australian AI companies 
could be an advantage, enabling companies to pivot quickly 
to changing market conditions and respond to early-stage 
AI technology developments.

Geographic hotspots and 
clusters are forming – AI companies 
like to be near other AI companies

We examined spatial clustering patterns of AI companies in 
Australian cities and found that they are tightly clustered. 
We identified eight clusters in Sydney, Melbourne, 
Brisbane, Adelaide and Perth, with more forming in other 
cities and regions. We found that 54% of AI companies 
were in one of the clusters and that the average distance 
from an AI company to its nearest neighbour AI company 
was 80 m. This suggests we are likely to see the emergence 
of place‑based AI hubs in Australia in the future.

Australia’s R&D sector 
has higher (and growing) rates 
of AI adoption compared to 
global averages

Australia has higher intensity of AI research than the global 
average. In Australia, 9.6% of all R&D publications in 2022 
referred to AI technologies compared to 7.2% globally. 
The rate of research publications has also been increasing. 
In 2023, AI R&D publishing by public and private sector 
organisations accounted for 9.6% of all R&D publishing 
which is up from 9.4% the year before, 4.9% in 2015 and 
2.5% in the year 2000. Since 2000, the Australian R&D 
sector has published 92,940 peer-reviewed books, book 
chapters, conference papers and journal papers about AI.



Australia creates much AI 
knowledge but relatively few 
commercial products

We found that Australia contributes 1.6% of global 
peer‑reviewed research publishing on AI. By comparison, 
we found that Australia contributes only 0.2% of global 
AI patent applications where the inventor resides within 
Australia. This shows that we are not converting our 
knowledge and understanding of cutting-edge AI science 
into commercial products at the same rate as other 
countries. Essentially, we give the world more knowledge 
about AI than commercial AI products. This sentiment 
was echoed in consultations with stakeholders across 
the Australian AI ecosystem.

Australia has a mature 
AI ecosystem with clear 
specialisations in AI technology 
and applications domains

Stakeholders across Australia’s AI ecosystem identified a 
desire to focus on specialised areas for AI development, 
rather than positioning Australia as a “jack of all” AI 
capabilities. Our analysis highlighted current areas of AI 
specialisation when it comes to specific AI technologies 
(e.g. self-organising maps, ant colony optimisation, gradient 
tree boosting) and application domains (e.g. optometry, 
livestock production and decision sciences). Beyond these 
R&D-driven strengths, stakeholders also highlighted 
opportunities for Australia to take a world-leading role in 
areas complementary to existing industry and geographical 
strengths, or AI technologies that emphasise Australia’s 
trusted brand.

Australia is recognised as a 
place for trusted and responsible 
AI innovation

The Australian brand is a key asset that can be leveraged 
when it comes to the development of safe and responsible 
AI. In addition to focusing on specialised areas of 
strategic advantage, stakeholders from the AI ecosystem 
identified the strength of Australia’s reputation as a 
comparatively trustworthy and appropriately regulated 
country. This global positioning, combined with Australia’s 
diverse population which is beneficial when developing 
AI technologies that are generalisable to diverse markets, 
could be used to position Australia as a global leader 
in the development and use of safe and responsible 
AI technologies.



Planned improvements

Future AI ecosystem reports

This is the second AI ecosystem report released by 
CSIRO’s National Artificial Intelligence Centre. It captures 
a broader range of source data compared to last year’s 
report. And there are plans to expand and improve the 
metrics captured and communicated in future reports. 
In this section we explore some of the main planned 
improvements for future versions.

1. Improved data on the AI workforce 
(workers and skills demand profile)

We know that most of Australia’s AI workforce is not in 
one of the AI companies we identified. They’re in banks, 
mining companies, software companies, supermarkets 
and a wide range of other companies that are developing 
AI capabilities to achieve improved business processes. 
At the current time, there is no formal classification of 
an AI worker in the Australian and New Zealand Standard 
Classification of Occupations (ANZSCO). This means we 
can’t get statistics from formal government sources 
(e.g. The Australian Bureau of Statistics) on the number 
of, and types of, AI workers in Australia. However, it may 
be possible to use alternative online data about AI jobs 
(e.g. job adverts). By meshing ANZSCO data from the census 
with online data, we could build a more detailed picture 
of Australia’s AI workforce.

2. Improved data on AI training 
and education providers

Australia’s schools, universities and technical colleges 
provide extensive education and training on AI-related 
topics. These include skills and knowledge areas such 
as machine learning, computer vision, natural language 
processing, robotics, mathematics, statistics, software 
engineering, computer coding and data science. It also 
covers skills and knowledge such as AI business strategy, 
technology foresight, government policy and planning, 
ethics, user experience/design, human-computer interfaces 
and organisational transformation. The AI training 
ecosystem is likely to be expanding and diversifying 
in Australia to meet the demand for these skills and 
knowledge. Future versions of this report could examine 
the size and structure of Australia’s AI training and 
education sector.

3. Analysis of AI adoption patterns 
and productivity impacts

There is an increasing body of evidence about productivity 
uplift associated with AI adoption. For example, a recent 
study by Stanford University (Brynjolfsson et al., 2023) 
found that customer support staff were able to solve 14% 
more complex customer questions per hour when they 
used generative AI chatbots. It was also found to increase 
customer satisfaction and the productivity gains were more 
than double for new and inexperienced staff. However, not 
all AI adoption stories are about success. There is much 
complexity and challenge about how AI is adopted, which 
tools are used and how they are applied (Hajkowicz and 
Whittle, 2023). Future versions of the AI Ecosystem report 
could explore the extent of adoption, patterns of adoption 
and productivity impacts.



References

Arundel A, Kabla I (1998) What percentage of innovations 
are patented? empirical estimates for European firms. 
Research Policy 27(2). DOI: 10.1016/S0048-7333(98)00033-X.

AIHW (2023) Culturally and linguistically diverse Australians. 
Australian Institute of Health and Welfare, Canberra. 
Retrieved from: https://www.aihw.gov.au/reports-data/
population-groups/cald-australians/overview.

Australian Bureau of Statistics (2023) Counts of Australian 
Businesses including entries and exits, June 2019 to June 
2023 (Catalogue Number 8165.0). Australian Bureau of 
Statistics, Canberra.

Australian Government (2023) Intergenerational report 
2023: Australia’s future to 2063. Treasury, Australian 
Government, Canberra.

Bratanova A, Pham H, Mason C, Hajkowicz S, Naughtin 
C, Schleiger E, Sanderson C, Chen C, Karimi S (2022) 
Differentiating artificial intelligence activity clusters 
in Australia. Technology in Society 71, DOI: 10.1016/j.
techsoc.2022.102104.

Brynjolfsson E, Li D, Raymond LR (2023) Generative AI at 
Work. National Bureau of Economic Research working 
paper 31161, DOI: 10.3386/w31161.

Cameron A (2022) Building an innovation cluster: 
Approaches and policies to stimulating new industry.

CSIRO Publishing. Brisbane, Australia.

Crawley C, Beynon M, Munday M (2013) Making location 
quotients more relevant as a policy aid in regional spatial 
analysis. Urban Studies 50(9): 1854–1869.

Crunchbase (2023a) Crunchbase database. Retrieved from: 
https://www.crunchbase.com/home (June–August 2023).

Crunchbase (2023b) What Industries are included in 
Crunchbase? Crunchbase Product Team post. Retrieved 
from: https://support.crunchbase.com/hc/en-us/
articles/360043146954-What-Industries-are-included-in-
Crunchbase- (12 August 2023).

CSIRO (2023) Australian AI Ecosystem Discoverability 
Platform Beta. Retrieved from: https://www.csiro.
au/en/research/technology-space/ai/ai-capabilities-
directory?start=0&count=12 (14 July 2023).

Datarock (2023) Home page. Retrieved from:
https://datarock.com.au/.

DISR (2023a) Artificial intelligence. Retrieved from: https://
www.industry.gov.au/science-technology-and-innovation/
technology/artificial-intelligence. Department of Industry, 
Science and Resources, Australian Government, Canberra.

DISR (2023b) Safe and responsible AI in Australia: Discussion 
paper, Department of Industry, Science and Resources, 
Australian Government, Canberra.

DiUS (2023) Datarock: ML powered drill core analysis 
software. Retrieved from: https://dius.com.au/case-studies/
dius-and-solve-geosolutions-extract-new-value-in-mining-
using-artificial-intelligence.

Fifth Quadrant (2022) Responsible AI Index 2022: Report. 
Fifth Quadrant, Sydney.

Goode K, Kim H M, Deng M (2023) Examining Singapore’s 
AI progress. Centre for Security and Emerging Technology, 
Georgetown University’s Walsh School of Foreign Service, 
Washington DC.

Hajkowicz S, Evans D, Cameron L, Trinh K, Chen H, 
Bratanova A, Mason C, Pham H, Travis S, Black R (2023a) 
The Geography of Australia’s Digital Industries: Digital 
technology industry clusters in Australia’s capital cities 
and regions. CSIRO and the Technology Council of 
Australia, Canberra.

Hajkowicz S, Sanderson C, Karimi S, Bratanova A, Naughtin 
C (2023b) Artificial intelligence adoption in the physical 
sciences, natural sciences, life sciences, social sciences and 
the arts and humanities: A bibliometric analysis of research 
publications from 1960–2021. Technology in Society 74, 
DOI: 10.1016/j.techsoc.2023.102260.



Hajkowicz S, Reeson A, Evans D, Bratanova A, Cameron L 
(2021) Industry Growth Opportunities: A technical report to 
support the Western Parkland City Economic Development 
Strategy. A research report for the New South Wales 
Government by CSIRO Data61 Insights, Australia.

Hajkowicz SA, Whittle J (2023) Yes, AI could help us 
fix the productivity slump – but it can’t fix everything. 
The Conversation, 22 June.

Harrison AI (2023) Home page. Retrieved from:
https://harrison.ai/.

Jones C M, Danaher L, Milne M R, Tang C, Seah J, Oakden-
Rayner L, Johnson A, Buchlak Q D, & Esmaili N (2021) 
Assessment of the effect of a comprehensive chest 
radiograph deep learning model on radiologist reports and 
patient outcomes: a real-world observational study. BMJ 
Open, 11(12), e052902.

Maslej N, Fattorini L, Brynjolfsson E, Etchemendy J, Ligett K 
et al. (2023) The AI Index 2023 Annual Report.

AI Index Steering Committee, Institute for Human-Centered 
AI, Stanford University, Stanford, California.

Muro M, Liu S (2021) The geography of AI: Which 
cities will drive the artificial intelligence revolution? 
Brookings Institute, Washington DC.

Nakazato S, Squicciarini M (2021) Artificial intelligence 
companies, goods and services: A trademark based analysis. 
OECD Science, Technology and Industry Working Paper 
2021/06, Organisation for Economic Cooperation and 
Development, Paris.

Observatory of Economic Complexity (2023) Economic 
Complexity Index. Organisation for Economic Cooperation 
and Development, Paris.

OECD (2023) Revealed technology advantage in selected 
fields. DOI: 10.1787/data-00673-en. Organisation for 
Economic Cooperation and Development, Paris.

Pitchbook (2023) Pitchbook database. Retrieved from: 
https://pitchbook.com/ (June-July 2023). Productivity 
Commission (2022) 5-year productivity inquiry: Australia’s 
data and digital dividend.

Australian Government, Productivity Commission, Canberra.

Reid A, O’Callaghan S, & Lu Y (2023) Implementing 
Australia’s AI Ethics Principles: A selection of Responsible 
AI practices and resources. National Artificial Intelligence 
Centre and Gradient Institute, Sydney.

Sapia (2023) Home page. Retrieved from: https://sapia.ai/.

Stanford HCAI (2023) Global Vibrancy Tool – Artificial 
Intelligence Index. Stanford University Human-Centered 
Artificial Intelligence. Retrieved from: https://aiindex.
stanford.edu/vibrancy/ (23 August 2023).

Swoop Aero (2023) Home page. Retrieved from:
https://swoop.aero/.

Toney A, Flagg M (2020) US Demand for AI-Related Talent. 
Center for Security and Emerging Technology (CSET), 
Georgetown University, Washington DC.

University of Toronto (2020) Canada’s AI ecosystem: 
government investment propels private sector growth, 
University of Toronto, Canada.

VAPAR (2023) Retrieved from: Home page.
https://www.vapar.co/.



Dr Stefan Hajkowicz
Chief Research Consultant
Analytics and Decision Sciences

The National AI Centre is 
building Australia’s responsible 
and inclusive AI future.

For further information

National AI Centre

1300 363 400
+61 3 9545 2176
naic@csiro.au
csiro.au/naic