Accelerating Australia’s

AI Agenda
A positive vision to make Australia a globally recognised
AI leader by 2028.
GPO Box 1472, Melbourne 3001
T 03 8664 2664
[Link]

© Copyright June 2025 Business Council of Australia, ABN 75 008 483 216.
All rights reserved. No part of this publication may be reproduced or used in any way
without acknowledgement to the Business Council of Australia.
Release date: June 2025
Contents
01 Australia’s AI opportunity.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 01
Giving us the power to do more. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 02
Our national productivity challenge. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 02
A mindset shift. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 03
Our call to action. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 03
02 List of actions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 05
03 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 07
Types of Al. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 08
A new economic opportunity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 09
Australia in the world of Al. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 09
A case for optimism and opportunity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
04 Vision and timeline. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
05 Defining Australia’s Al context. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
The productivity imperative. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Global competition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
A human-centric approach. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Leveraging Australian advantages.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
06 Australia’s Al Agenda. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Adoption.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Skills. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Regulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Infrastructure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
R&D and innovation.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
07 Conclusion.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Business Council of Australia i

ii Accelerating Australia’s AI Agenda
 01

Australia’s AI opportunity

Australia stands today at

the dawn of something truly
transformative. Driven by
incredible leaps in computing
power and data, AI has reached
a tipping point. What we do now
will determine whether Australia
is playing catch up for decades
or gains a competitive edge that
improves all Australian lives.

Business Council of Australia 1

 01 Australia’s AI opportunity

Giving us the power to do more Our national productivity

AI is already empowering us to do, challenge
see, and understand more as we go With this much opportunity at stake, we
about our lives, from Google Maps, must seize it with urgency as a national
to voice assistants, and email spam priority. This technological shift comes at
a critical moment. Long-term economic
filters. And the ways it can support
growth, the kind that lifts living standards
us are growing dramatically.
for everyone, doesn’t happen by accident.
AI is helping doctors diagnose diseases It hinges on our ability to innovate, to
earlier and more accurately, giving families invest wisely, and to quickly embrace new
time and hope. It is helping heavy industries technologies that allow us to do or make
operate more safely, reducing waste more without more effort or resources.
and creating new, high-value products. This is known as productivity, and it has
In education, AI is helping personalise languished in Australia for decades. The
learning experiences while giving teachers great leaps forward we have made in
time back to spend with students. AI is productivity have been a result of new
helping us manage our natural resources technologies – electricity, mechanisation,
more sustainably. It will underpin our computers. The next frontier is AI. Australia
economy and quality of life in the same must make big moves immediately to
way as electricity – an essential, seamlessly ensure we do not miss the opportunity.
integrated force empowering how we live
and work.1

1 Ng, A 2017, Why AI Is the New Electricity, Stanford Business ([Link]

2 Accelerating Australia’s AI Agenda

 01 Australia’s AI opportunity

A mindset shift OUR CALL TO ACTION

Like any new technology, there are risks
This paper outlines a clear, positive
which we must address head-on. But we
have been balancing risks with opportunity vision and practical actions to make
since the discovery of fire. What matters Australia a globally recognised AI
are the choices we make – how we develop leader by 2028. We must focus
our AI capabilities, how we deploy them, our efforts in six areas: adoption,
and how we ensure they serve our national skills, regulation, infrastructure,
interests and reflect our values. We could
data, and R&D. Within each area, we
choose to do this with a fear-first mindset,
but a brighter future will come from being
propose actions designed to foster
focused first on opportunity. innovation, attract investment,
unlock productivity gains, manage
risks responsibly, and make
Australia a more prosperous and
resilient nation.

Contents >

Business Council of Australia 3

4 Accelerating Australia’s AI Agenda
 02
List of actions

Action 1 Government as an exemplar: pioneering AI in public services .. . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Action 2 Boost funding for the National AI Centre (NAIC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Action 3 Create a National AI Skills Compact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Action 4 Reform AI skills development pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Action 5 Implement clear, risk-based regulations .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Action 6 Establish the Australian AI Safety Institute (AAISI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Action 7 Simplify data centre development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

Action 8 Reform copyright through the National AI Capability Plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Action 9 Expand frameworks for public-private data access and sharing .. . . . . . . . . . . . . . . . . . . . . . . . . 32

Action 10 Establish the Australian National Data Library (ANDL) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Action 11 Create secure private sector data collaboration frameworks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Action 12 Empower researchers and enhance research infrastructure .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

Action 13 Establish the AI Research Consortium (AIRC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

Action 14 Fix R&D settings to support AI development and adoption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Action 15 Create the AI Commercialisation Accelerator (AICA) .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Action 16 Establish a national AI missions program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Contents >

Business Council of Australia 5

6 Accelerating Australia’s AI Agenda
 03

Introduction

There are moments in history

when the future pivots. The
invention of the printing press,
the harnessing of steam power,
the dawn of the internet –
these were inflection points
that reshaped societies and
economies in ways previously
unimaginable.

Business Council of Australia 7

 03
Introduction

Types of AI
AI technology encompasses far more experiences. Natural language processing powers
than just chatbots like ChatGPT. translation services and document analysis.
Machine learning systems analyse Robotics systems in manufacturing work alongside
humans to handle precision or dangerous tasks.
medical images to detect diseases
AI also drives sophisticated decision support
earlier than human doctors can.
systems in fields ranging from finance to climate
Computer vision allows self-driving vehicles to science, processing vast datasets to identify
navigate complex environments. Recommendation patterns humans might miss.
engines personalise your streaming and shopping

8 Accelerating Australia’s AI Agenda

 03 Introduction

A new economic opportunity

After decades of integrating machine learning into
many platforms and services we use every day, the
convergence of unprecedented computing power,
the availability of vast datasets, and increasingly
sophisticated algorithms has supercharged AI
into a general-purpose technology.2 And this is
why it has massive economic potential. A general-
purpose technology is broadly applicable, creates
other innovations, and will continually improve.3
Automation technologies – which include AI –
could add $600 billion to our economy annually.4
AI is a pathway to reignite productivity growth,
sharpen our competitive edge on the global stage,
enhance the delivery of essential services like
healthcare and education, and develop innovative
solutions to complex national challenges. For
example, Telstra uses AI to analyse vast amounts
of data from weather stations, satellites, and
social media during bushfires. This allows them
to predict fire spread patterns and identify areas
at high risk. This information is then relayed to
emergency services, enabling them to deploy
resources more effectively and keep
communities safer.
However, we must be wise in how we harness this
power. The very power that makes AI so promising
also raises legitimate questions about how we use
it responsibly and safely. We cannot ignore the
concerns people have about the future of work,
the protection of personal data and privacy, the
ethical dilemmas posed by autonomous systems, Australia in the world of AI
and the risk of misuse by those who wish us harm.
Australia approaches this new era from a unique
Public trust is the bedrock upon which successful
position. We are blessed with significant strengths:
adoption rests. Without it, even the most beneficial
a highly educated and adaptable workforce, stable
technologies may falter. Both industry and
democratic institutions, abundant renewable
government have roles to play in strengthening
energy potential that can power the future, and
public trust.
world-class research institutions. Australia is also a
part of the powerful Five Eyes intelligence sharing
arrangement with the US, UK, Canada and
New Zealand, as well as AUKUS with the US and
UK, which includes a work stream on AI and
autonomy. Both of these security arrangements
give Australia privileged access to AI technology
sharing and development initiatives, boosting our
national security.

2 Buchanan, B and Imbrie, A 2022, The New Fire, MIT Press, p.2
3 Ben-Ishai, B, Dean, J, Manyika, J, Porat, R, Varian, H, Walker, K, 2024, AI and the Opportunity for Shared Prosperity: Lessons From the History of Technology and the Economy,
Google, pp.4-6 ([Link]
4 McKinsey & Company 2019, Australia’s automation opportunity, p.7 ([Link]
reigniting-productivity-and-inclusive-income-growth)

Business Council of Australia 9

 03 Introduction

These are powerful assets, but we must not in Australia align with our national interests, our
become complacent. We also face distinct economic needs, and our democratic values.
challenges. Compared to citizens in many The Australian Government’s yet-to-be written
other nations, Australians express a higher level National AI Capability Plan – slated for completion
of nervousness about AI’s impact. Our digital by end of 2025 – presents a golden opportunity to
infrastructure, while in some areas world-leading, get this right.6
needs significant acceleration to meet the
demands of widespread AI deployment in what is
now a global race. And Australia faces the perennial A case for optimism and opportunity
challenge of translating our research excellence The BCA’s view of AI is optimistic, but it is
into commercial success and global scale. not blind optimism. It adopts a positive,
We must acknowledge the fierce global opportunity‑focused perspective, grounded in
competition for AI leadership. This is not a the conviction that AI, guided by human
race Australia can afford to sit out. The window principles and smart policy, can be a powerful
of opportunity to shape the trajectory of AI force for good. We believe AI can augment
development and deployment is closing. We human capabilities, create new kinds of jobs,
must move with urgency and ambition to avoid and make existing ones more rewarding and less
becoming merely consumers of AI technology burdensome. But achieving this positive future
developed elsewhere. The UK’s AI Opportunities requires deliberate choices and concerted action.
Action Plan has succinctly articulated this as The ultimate goal here isn’t technological
an aim to become an ‘AI maker’ not merely supremacy. It’s about building a better future for all
an ‘AI taker.’ 5 This means fostering our own Australians, one where we have greater power and
domestic capabilities, attracting global talent and agency in our lives.
investment, and ensuring that the AI systems used

Contents >

5 UK Government 2025, AI Opportunities Action Plan ([Link]

6 Department of Industry, Science and Resources 2024, Developing a National AI Capability Plan ([Link]

10 Accelerating Australia’s AI Agenda

 04

Vision and timeline

Our vision is clear, ambitious, and

achievable: By 2028, Australia
will be a global leader in AI –
shaping trusted, transformative
technologies that power
economic growth, improve lives,
and build a more resilient nation.

Business Council of Australia 11

 04
Vision and timeline

Phased approach

The timeline is structured across

three years in a phased approach
to development:

ƒ Year 1: 1 July 2025 – 30 June 2026

ƒ Year 2: 1 July 2026 – 30 June 2027
ƒ Year 3: 1 July 2027 – 30 June 2028

All actions should start as soon

as possible in 2025, with a clear
statement of ambition from the
federal government. Completion
dates vary with each action’s
complexity and its dependency
on other actions.

YEAR 1 YEAR 2 YEAR 3

Establishing Capacity building Advancing
the essential and infrastructure collaboration
foundations development and optimising
data utilisation

12 Accelerating Australia’s AI Agenda

 04 Vision and timeline

Timeline

Year Phase Actions

YEAR 1 Establishing ƒ Action 1: 

Government as an exemplar: pioneering AI in
the essential public services
foundations ƒ Action 2: Boost funding for the National AI Centre (NAIC)
ƒ Action 3: Create a National AI Skills Compact
ƒ Action 5: Implement clear, risk-based regulations
ƒ Action 7: Simplify data centre development
ƒ Action 13: Establish the AI Research Consortium (AIRC)

Y EAR 2 Capacity building ƒ Action 4: 

Reform AI skills development pathways
and infrastructure ƒ Action 6: 
Establish the Australian AI Safety Institute (AAISI)
development
ƒ Action 8: 
Reform copyright through the National AI
Capability Plan
ƒ Action 14: Fix R&D settings to support AI development
and adoption
ƒ Action 15: Create the AI Commercialisation Accelerator (AICA)
ƒ Action 16: Establish a national AI missions program

YEAR 3 Advancing ƒ Action 9: 

Expand frameworks for public-private data access
collaboration and and sharing
optimising data ƒ Action 10: Establish the Australian National Data Library (ANDL)
utilisation
ƒ Action 11: 
Create secure private sector data collaboration
frameworks
ƒ Action 12: Empower researchers and enhance research
infrastructure

Contents >

Business Council of Australia 13

14 Accelerating Australia’s AI Agenda
 05

Defining Australia’s AI context

Several factors define Australia’s

context and will shape our
national approach to AI. These
include our geographic position
in the Asia-Pacific region, our
strong democratic institutions,
our skilled workforce and
research capabilities, and an
increasingly competitive global
technology landscape.

Business Council of Australia 15

 05Defining Australia’s AI context

The productivity imperative

right now. What if every call centre in Australia

Australia’s future economic prosperity
depends heavily on our ability to lift could handle twice as many customer inquiries
productivity.7 AI represents the single with half the wait time? What if your HR team could
process payroll and benefits in minutes instead
greatest opportunity to do this in a
of days? What if your supply chain could adapt to
generation, addressing some of our
disruptions before they impact your customers?
most prominent and enduring This isn’t futuristic thinking. This is possible today
productivity challenges.8 with basic machine learning applied to relatively
The most powerful AI applications aren’t boring tasks. And the productivity improvements
necessarily the flashy ones. They’re the ones that can be extraordinary.
solve real productivity problems people face

7 Business Council of Australia 2024, Australia’s flagging competitiveness and productivity: How to turn it around, BCA, p.5 ([Link]
competitiveness_and_productivity)
8 Productivity Commission 2024, Making the most of the AI opportunity, Research paper 1, AI uptake, productivity, and the role of government, Australian Government, p.3
([Link]

16 Accelerating Australia’s AI Agenda

 05 Defining Australia’s AI context

For example, Commonwealth Bank is using Imagine this effect scaled across all business,
AI to read, analyse and process customer government and research organisations. In an
documentation more quickly and accurately era of slowing productivity growth, AI offers a
than humans. This allows the bank to halve the powerful new engine. We must ensure that engine
time it takes to verify someone’s income when it is running smoothly and efficiently here
processes a loan. These tools are a win for staff in Australia.
as well, who can spend less time on repetitive
Australia needs practical frameworks, clear
tasks and focus more on having high-quality
guidelines, and accessible tools we can implement
conversations with customers.
on Monday morning. In the game of productivity,
When we optimise these daily pain points, we the winners won’t necessarily be those who build
don’t just save money and do more with less. the most sophisticated AI. The winners will also
We free our people to do what humans do best: be those who adopt and use AI most broadly, not
create, connect, and innovate. just in areas that are complex and meaningful, but
also those that are simple and mundane. It’s about
Studies have shown that AI enhances productivity
helping people unlock those incremental gains
and often helps close the gap between low-
that, when combined, create massive momentum.
and high-skilled workers.9 It boosts software
development productivity by 26 per cent.10
Generative AI for high-skilled work reduces
perceived mental demand by 45 per cent, and
perceived stress and difficulty by 58 per cent.11
Ninety per cent of small and medium businesses
using AI report more efficient operations.12

9 Stanford University Human-Centred Artificial Intelligence 2025, Artificial Intelligence Index Report 2025, p.18
10 Cui, Z, Demirer, M, Jaffe, S, Musolff, L, Peng, S and Salz, T 2025, The Effects of Generative AI on High-Skilled Work: Evidence from Three Field Experiments with Software
Developers ([Link]
11 Jaffe, S, Shah, N, Butler, J, Farach, A, Cambon, A, Hecht, B, Schwarz, M, and Teevan, J 2024, Generative AI in Real-World Workplaces, Microsoft Technical Report, p.9 (https://
[Link]/en-us/research/publication/generative-ai-in-real-world-workplaces/)
12 Salesforce 2024, Small & Medium Business Trends Report 6th Edition, p.6 ([Link]

Business Council of Australia 17

 05 Defining Australia’s AI context

Global competition Pioneers also get to set the global direction for AI
responsibility, with ethics and progress moving
Countries that recognise AI’s strategic importance together.14 Australia should learn from leading
are investing heavily. This is a race where early jurisdictions and apply best-practice policy
movers gain significant advantages – attracting settings to fast-track progress locally – this paper
talent, securing investment, setting standards, and identifies several overseas initiatives Australia
building competitive market positions. Australia should replicate.
cannot afford to be left behind.
The full integration of AI capabilities within R&D is
We must act decisively and strategically, fostering now a non-negotiable for global competitiveness
our own sovereign capabilities and attracting in advanced industries like biotech, materials
investment and talent from abroad. Australia is design, advanced electronics and defence
lagging in AI readiness compared to regional industries. Predictive foundation models are
peers, which underscores the need for rapidly accelerating the discovery of new
immediate action.13 materials, compounds and designs which are the
AI pioneers – such as the US, the UK, and lifeblood of these industries. Integrated AI systems
Singapore – have shown what’s possible through that continually learn over long periods and
steady investment, strong infrastructure, and a engage directly with humans in natural language
clear vision. They’ve built the basis for leadership are becoming core requirements of advanced
in adoption, innovation and talent development. defence systems.

13 In 2023, Australia ranked 28th in The Global AI Vibrancy Tool from the Institute for Human-Centered AI, Stanford University ([Link]
vibrancy-tool-2024). In 2024, Australia ranked 17th in The Global AI Index ([Link]
14 Schwaerzler, C, Carrasco, M, Daniel, C, Bollyky, B, Niwa, Y, Bharadwaj, A, Awad, A, Sargeant, R, Nawandhar, S and Kostikova, S 2024, The AI Maturity Matrix, BCG, p.4 (https://
[Link]/publications/2024/which-economies-are-ready-for-ai)

18 Accelerating Australia’s AI Agenda

 05 Defining Australia’s AI context

A human-centric approach
Technology must serve humanity. Our focus
must be on harnessing AI to benefit people
and businesses, not merely technological
advancement for its own sake. This means
prioritising applications that augment the
workforce, particularly in sectors facing chronic
skills shortages like healthcare and aged care,
allowing dedicated professionals to spend more
time on the human aspects of their roles. For a
nation with a relatively small workforce compared
to global giants, AI offers a way to maximise the
impact of our talented people.
Crucially, we must confront Australia’s AI anxiety
head-on. Australians, more than people in many
other countries, express nervousness about AI.15
Australians are the least optimistic in the world
about AI, and we struggle to believe its rewards
will outweigh the risks.16
We cannot dismiss these concerns. We must build
trust through transparency, robust governance
and regulatory frameworks, clear ethical
guidelines, and by actively demonstrating how
risks are being identified and managed.
We need a national conversation, grounded in
facts, that addresses fears while focusing on the
massive potential for positive transformation.
The narrative matters immensely.
Historical analysis of technological adoption
provides valuable context for understanding AI’s
potential impact on jobs. Technology typically
automates specific tasks rather than eliminating Studies indicate that 60 cent of the contemporary
entire occupations. Of the 271 occupations listed US workforce is employed in roles that did not
in the 1950 US Census, only one – the elevator exist 80 years ago.18 Current observations suggest
operator – was rendered obsolete due AI is aligning with this historical pattern, primarily
to automation.17 automating subsets of tasks within existing
occupations rather than causing job displacement.19
Since the mid-20th century, technology has
automated a growing number of human- When fear guides our response to AI, we risk over-
performed tasks, yet concurrently, new regulating. We stifle innovation, miss opportunities,
occupations have emerged. The proof is in and fall behind. That’s not effective leadership.
the pudding of our current unemployment Instead, let’s create smart rules targeting real risks.
rates, which sit lower than the days before When optimism leads, we create the space for
computerisation or electrification – both of AI to safely deliver its benefits for all Australians.
which were mistakenly decried as job-killers.

15 In 2023, Australia had the highest response globally to ‘Products and services using AI make me nervous’ at 69 per cent, compared to global average of 52 per cent. In 2024,
Australia was still in top 5 most nervous countries, clustered with the Anglosphere’s general high nervousness and low excitement about AI. Ipsos 2023, Global Views of AI
2023, p.8. Ipsos 2024, The Ipsos AI Monitor 2024, pp.15–17 ([Link]
16 Gillespie, N, Lockey, S, Ward, T, Macdade, A, & Hassed, G 2025, Trust, attitudes and use of artificial intelligence: A global study 2025, The University of Melbourne and KPMG
p.34
17 Bessen, J 2016, How computer automation affects occupations: technology, jobs, and skills, Boston University, p.5
18 Autor, D, Chin, C, Salomons, A, Seegmiller, B, 2022, New Frontiers: The Origins and Content of New Work, 1940–2018, National Bureau Of Economic Research, p.12
19 Agrawal, A, Gans, J, and Goldfarb, A. 2023, The Turing Transformation: Artificial intelligence, intelligence augmentation, and skill premiums, Brookings ([Link]
edu/articles/the-turing-transformation-artificial-intelligence-intelligence-augmentation-and-skill-premiums/)

Business Council of Australia 19

 05 Defining Australia’s AI context

Leveraging Australian advantages Our AI research community, while small, remains

globally competitive. In 2023, Australia had
Our political stability and strong democratic about 0.3 per cent of the global population
institutions provide a reliable foundation. We but contributed 1.6 per cent of all AI research –
boast a highly skilled and educated population overperforming by nearly five times.20 Australia is
and world-leading research institutions. By particularly strong in the design and development
strategically leveraging advantages, we can create of highly efficient AI that has low data, energy and
an environment that attracts global investment, processing requirements.
nurtures local talent, and builds a vibrant,
self‑sustaining AI ecosystem. We have vast land resources – especially in
contrast to our regional neighbours – critical for
We must identify our unique niches – areas like housing the data centres that power AI. Home
biotechnology, agriculture technology, field grown data centre developers such as AirTrunk
robotics, or mining automation – and double down are among the best in the world. Our potential for
on them. There are also areas that Australia can’t cost-competitive renewable energy is enormous,
afford not to support. The increasing medical cost offering a path to sustainable AI development.21
of an aging population is a significant political and Australia has long benefited from exporting
economic problem. Australia must implement AI in embedded energy to the world (such as coal
biotech and medical science to reduce costs and and gas). AI data centres powered by cheap and
improve treatments if we intend to maintain our reliable renewables are one way we can evolve
high standards of care. this sector.

Contents >

20 Hajkowicz S, Bratanova A, Schleiger E, Naughtin C 2023, Australia’s artificial intelligence ecosystem: Catalysing an AI industry, CSIRO, p.41
21 AirTrunk, Amazon Web Services, CDC Data Centres, Microsoft and NEXTDC 2025, Pre-Budget Submission 2025-26 Digital Infrastructure Priorities, p.4 ([Link]
[Link]/pre-budget-submissions/2025-26/view/642)

20 Accelerating Australia’s AI Agenda

 06

Australia’s AI agenda

Australia’s AI Agenda outlines

a vision and actions across
six key areas to become a
global AI leader by 2028,
boosting innovation and
national resilience.

Business Council of Australia 21

 06 Australia’s AI agenda

1. Adoption

The challenge with a strong capacity to spread adoption and

usage but weaker innovation capacity may be
The primary challenge for Australia is the rapid
more likely to sustain their success than those
pace of AI evolution and the risk of falling further
with the opposite.25 Credible estimates suggest
behind. Australia has already largely missed the
that increasing speed of adoption could increase
first major economic opportunity of AI, that of
generative AI’s addition to our economy from
generative AI and large language models (LLMs).
$45 to $115 billion annually by 2030.26
The era dominated by LLMs is already giving way
to a new paradigm: agentic AI, which operate with Australia should rapidly identify high-impact
a degree of independence, capable of setting innovations from abroad and adapt them to local
goals, making decisions, and executing complex conditions. By becoming world-class in how
tasks without constant human direction.22 This shift innovation is adopted and adapted, Australia
is the next major wave in AI.23 This demonstrates can punch above its weight economically and
the urgency of our situation. geopolitically. This ‘fast follower’ and ‘smart
adapter’ strategy can drive productivity gains,
One of the most important levers that
create new industries, enhance public services,
governments at all levels (federal, state and
and address unique national challenges.
territory) have to build confidence in AI is adoption
in the public sector. However, government uptake As 98 per cent of businesses in Australia, our
of AI in Australia has been slow. The federal overall adoption speed will be largely determined
government conducted a trial of generative AI by small and medium enterprises (SMEs). However,
from January to June 2024 and released Policy for recent findings paint a concerning picture: only
responsible use of AI in government in September one-third of SMEs have embraced AI technologies,
2024.24 More needs to be done to match the scale nearly a quarter struggle with implementation
and speed of the technology. expertise, and more than 40 per cent yet to
consider AI adoption.27 Despite governmental
support programs like AI Adopt Centres, the
The opportunity ecosystem supporting SMEs’ digital evolution
Australia should aim to be a ‘AI maker’ but the remains fragmented. In this complex environment,
AI opportunity is not only about inventing new SMEs find themselves at a crossroads between
things. Innovation capacity is very important to technological necessity and practical constraints.
a nation’s strategic and economic power, but Resolving this will be critical for Australia’s
speed of adoption and usage throughout an AI adoption.
economy can be even more important. This is
the process by which innovation moves from its
initial creation into widespread use, becoming a
standard product, a replicated engineering feat, or
a theory embraced across institutions. Countries

22 McKinsey & Company 2025, What is an AI agent? ([Link]

23 Savarese, S 2025, The Agentic AI Era: After the Dawn, Here’s What to Expect, Salesforce ([Link]
expect/)
24 Digital Transformation Agency 2024, Policy for responsible use of AI in government ([Link]
25 Ding, J 2023, The diffusion deficit in scientific and technological power: re-assessing China’s rise, Review of International Political Economy, 31(1), 173–198 ([Link]
0/09692290.2023.2173633)
26 Microsoft and Technology Council of Australia 2023, Australia’s Generative AI Opportunity, p.14 ([Link]
billion-annually-to-australias-economy-by-2030/)
27 Department of Industry, Science and Resources 2024, Exploring AI adoption in Australian businesses ([Link]
businesses)

22 Accelerating Australia’s AI Agenda

 06 Australia’s AI agenda

Action 1 Government as an exemplar:


2. Put AI leaders in charge to speed things up
pioneering AI in public services
To drive AI adoption, the Australian Public
Owner: Government Service should appoint Chief AI Officers
Timeframe: Complete by June 2026 (Year 1) (CAIOs) within two months and establish AI
Governance Boards within three. It may be
suitable to appoint one CAIO per ministerial
Industry is making significant strides in AI adoption portfolio or have fewer CAIOs covering
and innovation, demonstrating the technology’s multiple portfolios. These bodies will oversee
transformative potential. However, for Australia to AI use, ensuring adherence to governance
truly harness the power of AI across all sectors, the requirements such as compliance plans,
Australian Government must also spearhead AI updated AI policies, and maintained AI use case
adoption by becoming a sophisticated and ethical inventories. A CAIO Council should coordinate
user. This involves implementing AI in healthcare these efforts government-wide.
for improved diagnostics and efficiency, in
education for personalised learning and educator
support, and across government services to 3. Make sure government-used AI is safe for
enhance responsiveness and resource allocation. Australian businesses and people
AI also has great potential to review, simplify and Government AI must be trustworthy, secure,
improve complex regulation and policies across all and accountable. Agencies must develop
areas of government work. Such leadership will AI risk management policies and implement
demonstrate AI’s benefits, establish best practices, robust safety measures, addressing high-risk
build public trust, and encourage broader private AI use cases.
sector uptake.
To do so, the government should implement the
following three initiatives28:
Action 2 Boost funding for the National

AI Centre (NAIC)
1. Make it easier to use AI and get good value
Commonwealth agencies must identify
Owner: Government
and remove barriers to responsible AI use, Timeframe: Complete by June 2026 (Year 1)
developing agency-specific AI strategies
within six months. They should ensure data
compatibility for AI systems and share effective The NAIC was established in 2021 to accelerate
tools across government, while also focusing Australia’s AI industry and support AI adoption for
on recruiting, developing, and retaining small and medium businesses.29 The NAIC was
AI talent. funded for $21.6 million over four years from
2024-25. This must be boosted significantly to
reflect the urgency and scale of the challenge.
An expanded NAIC would be able to deliver
initiatives like AI implementation grants and
building on the AI Adopt Centres to provide
targeted financial incentives, technical assistance,
and expert consultancy through accessible
regional hubs. This will help SMEs identify relevant
AI solutions, overcome implementation barriers
such as data readiness, workforce training and
integrate AI into their business processes.

28 These are inspired by US executive memos: US Executive Office of the President 2025, Accelerating Federal Use of AI through Innovation, Governance, and Public Trust,
M-25-21 ([Link]
and US Executive Office of the President 2025, Driving Efficient Acquisition of Artificial Intelligence in Government, M-25-22 ([Link]
uploads/2025/02/[Link])
29 Department of Industry, Science and Resources, National Artificial Intelligence Centre ([Link]
artificial-intelligence-centre)

Business Council of Australia 23

 06 Australia’s AI agenda

2. Skills

The challenge The opportunity

The most sophisticated AI system is useless The prize for getting this right is immense: an
without people who know how to build, deploy, AI-skilled workforce gives Australian people
manage, and work alongside it. Right now, and businesses a fighting chance in a rapidly
Australia faces a critical challenge: ensuring our transforming world. These workers will be the
workforce, both current and future, possesses the drivers of higher national productivity and
necessary AI skills. enhanced global competitiveness. Businesses
equipped with AI-savvy employees will innovate
Australia needs to be training more data scientists
faster, operate more efficiently, and create
and machine learning engineers. It also needs
higher‑value products and services. Start-ups
to develop broad AI literacy across the entire
led by world class talent will be enormous wealth
economy. We need managers who embrace and
creators for this country. And a workforce ready to
drive AI adoption, workers who are comfortable
deploy AI at scale will draw global investment and
using AI tools in their daily jobs, and a public that
anchor high-value industries here in Australia.
engages confidently with AI’s societal impact.
By strategically developing and deploying AI to
This starts with basic IT literacy. In 2022, only
augment human capabilities, we can overcome
55 per cent of Year 6 students reached or
persistent skills gaps across our services sector, for
exceeded the proficiency standard, and
example in healthcare, education, and aged care,
only 46 per cent of Year 10 students.30 We need
improving service delivery and reducing burnout
to raise the general digital literacy of all students
among essential workers.
if we hope to have a future workforce that can
fully leverage AI.
Australia also needs the tradespeople that will
Action 3 Create a National AI Skills

build and maintain the digital infrastructure
Compact
required – primarily construction workers,
data centre operators, electricians and air Owner: Industry, academia and government
conditioning technicians. Timeframe: Complete by June 2026 (Year 1)
Australia faces stiff global competition for top
AI talent, and we need clear strategies to attract
Building Australia’s AI workforce will require shared
and retain the best minds.
commitment from government, the digital industry
and education providers. A National AI Skills
Compact would formalise this collaboration,
creating a structured partnership to scale talent
pipelines, align training with industry needs, and
expand access to AI careers. The Compact would
provide a clear mechanism for employers to signal
demand, shape curriculum, and support practical
learning pathways, while giving educators up-to-
date insight into the skills industry needs most.
Modelled on the NSW Digital Skills and Workforce
Compact, this national initiative would encourage

30 ACARA 2022, National Assessment Program: ICT Literacy 2022 Technical Report, p.8

24 Accelerating Australia’s AI Agenda

 06 Australia’s AI agenda

employers to commit to tangible actions –

including offering applied learning opportunities, ƒ To drive uptake and build credibility, members
recognising high-quality microcredentials, and of the proposed National AI Skills Compact
broadening recruitment beyond traditional should formally recognise industry-approved
pipelines. By anchoring AI workforce development microcredentials in recruitment. This includes
in real-world demand, the Compact would help allocating at least 20 per cent of entry-level
ensure training delivers job-ready graduates and AI roles to candidates from non-traditional
supports inclusive, responsive skills growth.31 pathways and committing to consider all
applications from graduates of approved
There would be two specific streams: Institute of Applied Technology-style programs.
These actions will broaden access to high-
value careers and help deliver a more diverse,
1. Create a national AI apprenticeship model
job‑ready AI workforce.
ƒ Develop a national apprenticeship model for
AI and digital infrastructure roles. This would
combine formal training with structured,
paid work placements in high-demand areas
like data engineering, machine learning Action 4 Reform AI skills development
deployment, and AI-related trades such as pathways
data centre operations and network systems.
Owner: Industry, academia and government
Drawing on lessons from Germany32 and
Timeframe: Complete by June 2027 (Year 2)
Singapore33, the model would offer a practical
pathway for school leavers, career changers,
and vocational learners to gain hands-on
experience while earning industry-recognised
AI literacy must start early and be embedded at
qualifications. every stage of learning. Foundational concepts
and practical skills should be integrated into
ƒ A nationally coordinated approach to AI school curricula from primary through to secondary
apprenticeships would ensure consistency in levels. At university, we need more specialist AI
design and delivery, while giving employers degrees. But just as importantly, AI modules must
flexibility to tailor on-the-job learning to
be embedded across all disciplines, so graduates
evolving business needs. This would help build
in law, medicine, the arts and beyond understand
a more diverse, work-ready AI talent pipeline
how to critically engage with AI tools in their fields.
and strengthen the link between training and
Vocational training courses must also embed AI
workforce demand.
skills to prepare learners for digitally enabled jobs
across the trades and technical sectors.
2. Scale up industry-led AI microcredentials Equally critical is ensuring the existing workforce
ƒ Australia should replicate NSW’s Institute of isn’t left behind. We need accessible, high-quality
Applied Technology Digital model nationally to upskilling and retraining options – including
deliver fast, flexible microcredentials in AI and short, non-accredited courses – delivered flexibly
related fields.34 These short, targeted courses through TAFEs, universities and private providers.
are essential for upskilling and reskilling,
Government, industry and accrediting bodies should
particularly for mid-career workers and those
work with the tertiary sector to co-design curricula,
looking to switch into a digital career. The
identify relevant use cases by discipline, and explore
effectiveness of microcredentials, however,
incentive mechanisms to support implementation.
depends on deep industry involvement. All AI
Embedding AI literacy in generalist degrees and
microcredentials must be co-designed and
endorsed by leading employers, reflecting real
relevant VET courses will help ensure Australia’s
workforce needs and emerging technologies. future workforce is prepared not just to adapt to AI,
but to actively shape its responsible use.

31 NSW Government, NSW Digital Skills and Workforce Compact ([Link]

32 Germany’s dual vocational training system combines on-the-job training with classroom instruction. Trainees typically spend part of the week at a company and the
remainder at a vocational school. Federal Institute for Vocational Education and Training 2020, Young people study in the company and at school ([Link]
en/[Link])
33 Singapore’s AI Apprenticeship Programme (AIAP) is a nine-month, full-time program designed to develop local AI talent. Participants engage in deep-skilling and work on
real‑world AI projects across various industries, including healthcare, insurance, and government agencies. AI Singapore 2025, Growing our own Timber ([Link] )
34 NSW Government, IAT Digital: The new way to get digital skills ([Link]

Business Council of Australia 25

 06 Australia’s AI agenda

3. Regulation

The challenge The opportunity

Regulation in the age of AI is a tightrope walk. We must remember that AI is already regulated –
We need rules that build public trust and protect perhaps 80 per cent of potential issues are solved
citizens from genuine harm, such as deepfakes by applying existing legislation, such as Australian
manipulating opinions, biased algorithms Consumer Law and sector based regulations.37
perpetuating inequality, or breaches of Australia needs to enforce existing laws and
data privacy. amend legislation where edge cases are identified.
Effective AI regulation starts from the premise that Getting regulation right is also a source of
the application and use of AI is overwhelmingly competitive advantage. A clear, predictable, and
positive, but that risks at the edges need to be trusted regulatory environment makes Australia
mitigated. We must avoid strangling the very a more attractive place for AI investment and
innovation we seek to foster. Overly burdensome, deployment. Businesses can operate with
ambiguous, or premature regulations throttle greater certainty, and consumers can adopt AI
experimentation, deter investment, and put technologies with more confidence. By developing
Australian businesses at a disadvantage. a reputation for responsible AI governance,
Australia can lead globally, influencing international
That is what we are seeing in the EU. In 2024,
standards and attracting businesses that prioritise
the Draghi report found that the EU’s regulatory
ethical practices. Smart regulation fosters
burden hindered the EU’s competitiveness and
the social license necessary for AI to flourish.
performance in critical areas like AI adoption,
Conversely, as long as the regulatory question is
cloud computing capabilities, and digital
unsettled, business investment will stagnate and
infrastructure investment. It called for a
people will remain more hesitant about AI.
reduction in this regulatory burden to rekindle
economic dynamism.35 We must also acknowledge that while some new
risks may arise, other risks may be reduced. For
The EU AI Act’s approval processes and third-
instance, self-driving cars are can vastly reduce
party audits for high-risk systems add months to
accidents on roads. Despite the focus on their
product launches. Meta’s Llama 4 AI models, for
errors, they eliminate the much higher level of
instance, were delayed by 18 months in the EU
human driving errors. Our resources companies
compared to the US. Apple Intelligence faced
are using AI to improve safety outcomes for
similar setbacks due to EU’s Digital Markets Act.36
Australian workers in mines. BHP has rolled
This is an important reminder that it’s not just ‘AI
out driverless trucks in their mines, including
regulation’ – Australia should maintain awareness
Goonyella Riverside Mine in Queensland’s Bowen
of our broader regulatory environment, and how
Basin in Barada Barna country. With each truck
this might inhibit AI.
fitted with lasers front and back, obstacles
Finding the right balance, ensuring our approach are flagged instantly, causing the vehicle to
aligns sensibly with international partners, and stop automatically.
addressing existing legislative challenges (like
how AI interacts with current tax or privacy laws)
is a complex but critical task. The perception that
regulation is inherently ‘good’ or ‘bad’ is unhelpful,
we need regulation that enables the trust and
safety that allows innovation to flourish.
35 Draghi, M 2024, The Future of European Competitiveness—A Competitiveness Strategy for Europe ([Link]
36 Jackson, F 2025, EU AI Rules Delay Tech Rollouts, But Civil Societies Say Safety Comes First, TechRepublic ([Link]
civil-groups/)
37 Select Committee On Adopting Artificial Intelligence (AI) Senate Committee, 21 May 2024, Official Committee Hansard, p 22

26 Accelerating Australia’s AI Agenda

 06 Australia’s AI agenda

address AI-related risks within their areas. To assist

Action 5 Implement clear, risk-based

in the transition, Australia would also benefit from
regulations
increasing regulatory capacity.
Owner: Government
Framework legislation would ensure different
Timeframe: Complete by June 2026 (Year 1)
regulators can conduct assessments consistently
and apply common guardrails across diverse
sectors, promoting compliance and easing
The BCA’s submission to the government’s 2024
implementation burdens. Framework legislation
Proposal for Mandatory Guardrails in High-Risk
should be accompanied by shared definitions and
Settings recommended that government should
metrics for assessing AI risk.
conduct a transparent and thorough review of
current legislation to identify any AI-related gaps.38 However, we must avoid what we have seen
The BCA believes this has not yet been adequately transpire in the EU. The EU AI Act nominally
completed by government. establishes a risk-based framework, classifying
AI systems according to potential harm, but its
Only if this process reveals clear and compelling
practical implementation deviates significantly by
evidence of legislative shortcomings would
targeting AI at the foundational model level, and
dedicated AI ‘framework legislation’ be an
specifically general-purpose AI (GPAI) models,
appropriate response.
and even more stringently those deemed to
Such a regulatory framework should be risk- pose ‘systemic risk’. This shift from regulating the
based and proportionate, focusing scrutiny high-risk application to regulating the underlying
and safeguards on self-identified high-risk AI technology imposes substantial burdens and is
applications (such as in critical infrastructure, something Australia must avoid.
healthcare and law enforcement) that are aligned
An EU-style approach would also be difficult to
to international standards, and where the potential
implement and impose overly broad and rigid
for harm is significant. Applications identified
obligations, slowing the adoption of AI across
as low-risk should not be subject to regulatory
Australian sectors. Overregulating early-stage
oversight. Prioritising regulatory certainty is key
technologies can entrench incumbents and
– this means clear definitions (what constitutes
disadvantage SMEs and start-ups. Australia
‘high-risk’ AI?) and practical guidelines.
already faces strong global competition in AI
Rather than reinventing the wheel, framework development, and so overlaying a burdensome
legislation would empower our existing regulators regulatory regime would put local innovators at
– who already possess deep expertise in specific a structural disadvantage.
products, industries, and professions – by
If Australia regulates too broadly – if we act out
providing them with the principles and tools to
of fear – we risk shutting down the very kind of
experimentation that could help us solve the
problems we’re trying to address in the first place.
Australia could box itself into a future that’s less
innovative and less safe.39

38 Business Council of Australia 2024, Submission to the proposals paper for mandatory guardrails for AI in high-risk settings, BCA, pp.15–16
39 Bryan, K and Teodoridis, F 2024, Balancing market innovation incentives and regulation in AI: Challenges and opportunities, Brookings ([Link]
balancing-market-innovation-incentives-and-regulation-in-ai-challenges-and-opportunities/)

Business Council of Australia 27

 06 Australia’s AI agenda

It would share its findings transparently with

Action 6  stablish the Australian AI Safety
E
regulators, industry, and the public, enhancing
Institute (AAISI)
accountability. To be effective, AAISI needs
Owner: Industry, academia and government statutory independence, and secure funding
Timeframe: Complete by June 2026 (Year 2) from both government and industry as a
public‑private partnership.
Building a safer future with AI is also an economic
Trust requires verification. An independent,
opportunity. By co-developing these standards
respected body – the AAISI – would be focused on
and evaluation tools, AAISI would enable a market
the technical aspects of AI safety and reliability.40
for AI assurance and auditing services, allowing
The AAISI would also partner on, and be informed
businesses to demonstrate the trustworthiness
by, interdisciplinary work on AI ethics.
of their systems. Australia has a chance to lead
Crucially, AAISI would not be a regulator. Its role by building an economic sector focused on the
would be to collaborate internationally to develop provision of AI safety services and products.
technical standards, create robust methodologies An AAISI would be at the heart of that effort. With
for evaluating AI systems (testing for bias, security the AAISI, Australia can grow an entire ecosystem
and accuracy) and conduct safety research on of new AI safety companies, jobs, and exports.
frontier AI capabilities.41 Establishing an AAISI
means Australia will be able to work more closely
with other countries with AI Safety Institutes, such
as the US, UK, Japan, South Korea, Canada, France
and Singapore.

40 The BCA previously spoke favourably of establishing an Australian AI Safety Institute – see Business Council of Australia 2024, Submission to the proposals paper for
mandatory guardrails for AI in high-risk settings, BCA, pp.13, 16
41 Australia committed to establishing an AI Safety Institute under the 2023 Bletchley and 2024 Seoul declarations.

28 Accelerating Australia’s AI Agenda

 06 Australia’s AI agenda

4. Infrastructure

The challenge But infrastructure can’t be built overnight, and

the countries that build quickest will be better
Data centres serve essential functions. Our critical
placed to win the global race. ‘Speed-to-build’
services – healthcare, transport, emergency
is a critical determinate of where AI investment
response – rely on them.
is going. At the moment, Australia is a slow and
Data centre demand is being driven by the expensive place to build, with complex planning
increasing uptake of cloud services as our and approval processes for major infrastructure
economy continues to digitally transform. There projects. Globally, two months is becoming
is still enormous growth to be realised by moving best practice.
from on-premises servers to energy efficient
data centres.
The opportunity
Adding the expected uptake of AI services
by Australian businesses, demand for data By 2030, global capital expenditure on data
centre capacity will only grow. The data centre centres for AI processing loads will reach
deployable capacity here in Australia is projected US$5.2 trillion.46 We have abundant land, political
to more than double from 1,350 megawatts in stability, and some of the world’s best renewable
2024 up to 3,100 megawatts by 2030.42 The energy resources in solar and wind – if we can
Australian Energy Market Operator (AEMO) harness them quickly enough. This combination
estimates that data centres are currently about makes Australia a highly attractive location for
1.7 per cent of total consumption of the national building the large-scale, green data centres
electricity market (NEM)43, rising to five per cent by the world needs, positioning us as a key AI
2033-34 (under an aggressive data centre growth infrastructure hub for the Asia-Pacific region.47
scenario), and not rising higher.44 This would also support a range of other complex
computationally-intensive industries and activities
However, it’s important to understand that
such as climate modelling, genomic analysis,
aggregating compute and storage into data
drug discovery, computational fluid dynamics
centres means energy usage can be optimised.
modelling, seismic processing, and visual
If Australian businesses switched back today to
effects rendering.
using on-premise servers instead of hyperscale
cloud providers, it would consume an estimated Building this capacity not only supports our
67 per cent more energy.45 domestic AI ambitions but also creates significant
economic activity through construction, new jobs,
Australia faces a global race for this infrastructure.
and attracting further tech investment. It’s about
Investment capital and cutting-edge hardware (like
building the digital infrastructure for a 21st-century
advanced semiconductor chips, potentially affected
economy, for both AI and cloud services.
by global supply chain dynamics and regulations
like US export controls) are fiercely contested.

42 Mandala 2025, Empowering Australia’s Digital Future, p.6 ([Link]

43 3 TWh of 181.4 TWh total NEM - Australian Energy Market Operator 2025, National Electricity Market Factsheet ([Link]
national-electricity-market-fact-sheet)
44 Australian Energy Market Operator 2024, Electricity Statement of Opportunities 2024, p.35
45 Mandala 2025, Empowering Australia’s Digital Future, p.23 ([Link]
46 McKinsey & Company 2025, The cost of compute: A $7 trillion race to scale data centers, McKinsey Quarterly, p.2 ([Link]
and-telecommunications/our-insights/the-cost-of-compute-a-7-trillion-dollar-race-to-scale-data-centers#/)
47 Mandala 2025, Empowering Australia’s Digital Future, p.18 ([Link]

Business Council of Australia 29

 06 Australia’s AI agenda

For Australia, sustainability will be integral to

Action 7 Simplify data centre

our AI infrastructure plan. Powering AI with clean
development
energy is both an environmental necessity and
an economic imperative. Studies estimate that Owner: Government
AI could help mitigate 5 to 10 per cent of global Timeframe: Complete by June 2026 (Year 1)
greenhouse gas emissions by 2030, equivalent to
the total annual emissions of the EU.48
Australia needs to cut the red tape to speed up
AI is also critical to making renewable energy
data centre development. We need to collaborate
work. Renewables require complex balancing
across federal, state and territory and local
of supply and demand, weather forecasting for
governments to significantly accelerate planning
sun and wind, and energy storage management
and approval processes, providing clarity and
across various battery technologies. Australia must
speed for investors. Governments must take
accelerate our transition to renewable energy
immediate steps to create transparent timelines for
sources, and as renewables become a larger
approval processes.
proportion of our energy mix, and the system
becomes more complex, AI becomes necessary Establishing a dedicated public-private council,
to manage and optimise the grid. bringing together government agencies,
infrastructure providers, and tech companies,
would facilitate this coordination and troubleshoot
bottlenecks. Addressing the enabling
infrastructure for the future – particularly power
grid upgrades and sustainable water solutions –
must happen in parallel.

48 BCG and Google 2023, Accelerating Climate Action with AI, p.2 ([Link]/gumdrop/sustainability/[Link])

30 Accelerating Australia’s AI Agenda

 06 Australia’s AI agenda

5. Data

The challenge of its AI Opportunities Action Plan, the UK has

proposed a TDM exception that would include
Australia ranks among the very worst in the
an opt-out mechanism for rightsholders.52 This
OECD for data availability, data accessibility and
represents a shift towards legal clarity and a
government support for data re-use.49 Improving
pro‑innovation stance. It acknowledges that
this and facilitating better data usage will help
uncertain or overly restrictive copyright rules
realise productivity gains.50
can hold back AI research, investment, and
Data is the lifeblood of AI. Without access to competitiveness.
high-quality, diverse datasets, even the most
sophisticated algorithms cannot learn effectively.
However, much of the potentially valuable data The opportunity
in Australia remains locked away in silos – held
If we can overcome these challenges, the
by different government agencies, private
rewards are substantial. Open data ecosystems
companies, or research institutions – constrained
fuel AI innovation. Governments should facilitate
by technical barriers, commercial sensitivities, or
responsible data-sharing between public and
privacy concerns.
private sectors to accelerate AI development. The
Striking the right balance between enabling data Australian Government has already made good
access for innovation and rigorously protecting progress in this area. In 2022, the Data Availability
individual privacy, along with the complex issue and Transparency Act (DAT Act) was enacted,
of copyright and AI training, is one of the most creating a legislative framework for the secure and
significant challenges we face. Mishandling efficient sharing of public sector data. Curating
protected data is a surefire way to erode trust and high-quality, diverse, and representative Australian
support from the public. datasets can become a significant competitive
advantage, allowing us to train AI models that are
Even the EU has acknowledged the critical
uniquely attuned to our specific context, needs,
importance of access to information for
and values. To facilitate this, the UK has agreed
innovation. Under the Directive on Copyright in
to the creation of a National Data Library (NDL)
the Digital Single Market (Article 3 and 4), the EU
unlock data assets and boost the economy.53 They
has implemented a text and data mining (TDM)
are rapidly identifying and releasing the top five
exception.51 This allows both non-commercial
high-impact government datasets to demonstrate
and commercial entities to carry out data mining
the power of publicly owned data for AI training.
without needing to seek prior permission from
rightsholders, provided that lawful access to the Responsibly opening up access to anonymised
content exists. Article 4 contains a crucial opt‑out or aggregated public sector datasets, and
mechanism for rightsholders where they can creating secure mechanisms for private sector
expressly reserve their rights and exclude their data sharing, would fuel a vibrant ecosystem of
works from the scope of this exception. AI startups and researchers, driving innovation
across the economy. Government itself can lead
The UK has also recognised the economic and
by example, using data more effectively to improve
societal benefits of AI-fuelled innovation. As part
its own services and decision-making.

49 Organisation for Economic Cooperation and Development 2023, 2023 OECD Open, Useful and Re-usable data (OURdata) Index: Results and Key Findings, OECD Public
Governance Policy Papers, p.14 ([Link]
50 Productivity Commission 2024, Senate Select Committee on Adopting Artificial Intelligence (AI) Productivity Commission submission, p.6 ([Link]
supporting/adopting-artificial-intelligence)
51 European Union 2019, Directive (EU) 2019/790 of the European Parliament and of the Council of 17 April 2019 on copyright and related rights in the Digital Single Market and
amending Directives 96/9/EC and 2001/29/EC ([Link]
52 UK Government 2025, AI Opportunities Action Plan ([Link]
53 UK Government 2025, AI Opportunities Action Plan ([Link]

Business Council of Australia 31

 06 Australia’s AI agenda

Action 8  eform copyright through the

R Action 10  stablish the Australian National
E
National AI Capability Plan Data Library (ANDL)
Owner: Government Owner: Government
Timeframe: Complete by June 2027 (Year 2) Timeframe: Complete by June 2028 (Year 3)

The government’s attempt at copyright reform Based on the UK’s National Data Library (NDL),
and AI has been focused on transparency.54 Australia should establish a similar initiative to
Conducted through the Attorney-General’s serve as enabling infrastructure, designed to
Department, it takes a legalistic view focused on unlock the full potential of public-sector data by
potential harms, rather than enabling opportunity. facilitating secure, streamlined, and scalable
Government should shift this work out of the access to linked datasets.
Attorney-General’s Department and into the work
By breaking down existing data silos and
of the Department of Industry Science and
overcoming fragmentation, an Australian NDL
Resources on the National AI Capability Plan to
could drive AI innovation, transform public service
create a text and data mining exception to speed
delivery, and empower researchers and industry
up progress and emphasise the opportunity, as
to develop solutions tailored to Australia’s unique
has been done in the EU, US and will be done in
context and challenges. The core purposes
the UK. Reform is critical and work will go offshore
should be augmenting the Office of the National
if we don’t fix it.
Data Commissioner, removing systemic barriers
to data access and use, and ensuring public
data becomes a powerful, accessible resource
Action 9  xpand frameworks for
E for evidence-based policymaking, research
public‑private data access breakthroughs, and economic growth.
and sharing
To realise this vision, Australia’s NDL development
Owner: Government should be focused on impactful use cases aligned
Timeframe: Complete by June 2028 (Year 3) with national priorities rather than being purely
technology-driven. It should adopt a federated
model where data remains decentralised but
Australia needs to reconceive government data as accessible through common infrastructure,
a valuable sovereign resource and establish clear, governance, and standards.
modernised frameworks. These should build upon
existing initiatives like the DATA Scheme55 but expand
its scope and streamline its processes to significantly
increase the availability of public sector data for
research and innovation purposes, always under strict
ethical and privacy safeguards. Prioritising economic
growth and public benefit should be key objectives.
Similar to the UK’s ‘open by default’ policy, public
sector data should be published proactively
and made easily accessible for anyone to use,
modify, and share, unless there is a clear and
lawful justification to withhold it. The policy applies
broadly to non-personal data held by public
sector bodies. This includes data like government
spending, contracts, procurement data, ministerial
meetings, live traffic data, maps, land use, air
quality, weather, flood risks, and energy usage.

54 Attorney-General’s Department 2023, Copyright and Artificial Intelligence Reference Group (CAIRG) ([Link]
artificial-intelligence-reference-group-cairg)
55 Office of the National Data Commissioner 2024, Introducing the DATA Scheme ([Link]

32 Accelerating Australia’s AI Agenda

 06 Australia’s AI agenda

Potentially employing this sandbox model,

Action 11  reate secure private sector data
C
Australia should also support the creation of
collaboration frameworks
secure, trusted data collaboration platforms in the
Owner: Industry and government private sector. This could involve sector-specific
Timeframe: Complete by June 2028 (Year 3) ‘data trusts’ (for example, for health or agricultural
data) or cross-sectoral platforms inspired by
international examples like the EU’s Gaia-X57
Innovation often requires experimentation in safe and iSHARE58 and Dutch JoinData59 initiatives.
environments. We should actively implement These platforms would enable multiple parties
‘operational sandboxes’ for data-driven AI projects, to securely pool or access data for analysis and
allowing businesses and researchers to test new model development, using privacy-enhancing
ideas under regulatory supervision.56 technologies and robust governance to protect
confidentiality and commercial sensitivities.

56 Similar to what has already be established with the Consumer Data Right sandbox: Australian Government, Consumer Data Right Sandbox ([Link]
providers/participant-tooling/consumer-data-right-sandbox)
57 Gaia-X 2023, About Gaia-X ([Link]
58 iSHARE Foundation, iSHARE Trust Framework ([Link]
59 JoinData 2025, We are JoinData: an independent platform focused on farmers ([Link]

Business Council of Australia 33

 06 Australia’s AI agenda

6. R&D and innovation

The challenge models, giving it the “answers” needed to learn

how sequence determines structure.61
Becoming a global leader in AI requires a dynamic,
well-funded R&D ecosystem capable of generating A persistent issue in Australia is bridging the
foundational AI knowledge, translating discoveries ‘valley of death’ between research breakthroughs
into real-world applications, and nurturing the and commercial success – from January to June
next generation of innovators. The University of 2023, Australia contributed 1.6 per cent of the
Adelaide and Commonwealth Bank have already world’s AI research but only 0.24 per cent of AI
entered into a five year partnership to advanced patent applications.62
Australia’s AI research capability.60
Along with R&D that develops better AI, we must
The opportunity
harness AI’s potential to improve R&D across all
disciplines, such as biotechnology, engineering AI is more than a field of study in itself. It’s
and physics. This is essential if Australia is to becoming a universal tool for accelerating
remain a world-class research destination. discovery across virtually all domains of
science and innovation. AI has helped develop
Australia faces challenges here: ensuring sufficient
new medicines (like DeepMind’s AlphaFold
and consistent funding for both long-term
revolutionising protein folding prediction), design
research and more applied, industry‑focused
novel materials and modelling complex climate
development; equipping researchers across all
systems.63 AI has also driven major new scientific
scientific disciplines with the AI skills needed to
discoveries in battery optimisation, chemical
leverage its power in their own fields; ensuring
synthesis, and nuclear fusion reactors.64
researchers have practical access to powerful
computing resources essential for innovative Strategic investment in AI R&D can position
AI work; and overcoming fragmented funding Australia at the forefront of these AI-driven
approaches that can suffocate ambitious, scientific frontiers, creating high-value industries
large‑scale projects. and attracting world-class talent. A thriving
R&D ecosystem acts as a magnet, drawing in
High-quality data are particularly important.
investment and creating a virtuous cycle of
Consider AlphaFold, a breakthrough AI technology
innovation and economic growth. For example,
that solved the 50-year-old protein folding
Australia is a world leader in biotech, which is
problem, enabling scientists to predict a protein’s
also a relatively cheap and renewable industry for
3D structure from its amino acid sequence with
investment. It would make new companies and
unprecedented accuracy, revolutionising biology,
new jobs in Australian, and the breakthroughs in
drug discovery, and our understanding of life itself.
this area would be some of the most profound
This success was reliant on the Protein Data Bank
in human history.
– carefully annotated and triaged protein structure
data that research partners all over the world
collected for 60 years. This served as essential
training data for AlphaFold’s machine learning

60 Commonwealth Bank of Australia 2024, Five-year partnership to boost foundational AI research in Australia ([Link]
[Link])
61 Australian Research Data Commons, Making Data FAIR ([Link]
62 Hajkowicz S, Bratanova A, Schleiger E, Naughtin C 2023, Australia’s artificial intelligence ecosystem: Catalysing an AI industry, CSIRO, p.41
63 Tony Blair Institute 2025, A New National Purpose: Accelerating UK Science in the Age of AI, p.13 ([Link]
purpose-accelerating-uk-science-in-the-age-of-ai)
64 Wang, H, Fu T, Du Y, Gao, W, Huang K, Liu, Z, Chandak, P, Liu, S, Van Katwyk, P, Deac, A 2023, Scientific discovery in the age of artificial intelligence, Nature, 620(7972):47–60, p.48

34 Accelerating Australia’s AI Agenda

 06 Australia’s AI agenda

Action 12  mpower researchers and enhance

E Action 13  stablish the AI Research
E
research infrastructure Consortium (AIRC)
Owner: Industry, academia and government Owner: Industry and academia
Timeframe: Complete by June 2028 (Year 3) Timeframe: Complete by June 2026 (Year 1)

Alongside skills, researchers need the right tools Australia should establish a national AI Research
to meet the scale and speed of scientific change. Consortium (AIRC) – a dedicated research hub
This requires strategic investment in creating where businesses and universities co-locate to
AI-ready, high-quality research datasets. We pursue foundational, pre-competitive AI research.
should require publicly funded research data The AIRC would focus on how AI can drive
to be digitally recorded, well-documented, and real‑world productivity gains and economic
made accessible where appropriate under growth across all industries. By bringing together
FAIR (findable, accessible, interoperable and leading researchers and industry experts, the AIRC
reusable) principles.65 would help unlock the broad application of AI
technologies to solve shared challenges that are
We must also ensure Australian researchers have
too complex, risky, or resource-intensive for any
reliable and affordable access to the national
single organisation to address alone.
high‑performance computing (HPC) resources
needed to train and run sophisticated AI models. Modelled on world-leading initiatives like IMEC,
Investing in our research infrastructure is investing which has transformed Belgium into a global
in our future discoveries. leader in chip and nanoelectronics R&D, the AIRC
would focus on developing core AI capabilities
To achieve this, Australia should institute a
and non-proprietary technologies.66 These
program similar to Singapore’s ‘AI for Science’.
shared building blocks – such as new machine
Singapore is committing AUD $142 million to
learning architectures, trustworthy AI systems, and
harness AI in science, with the goal of increasing
advanced data infrastructure – would underpin a
research efficiency and unlocking new discoveries.
broad range of future commercial applications.
Critically, the investment will foster close
partnerships between AI specialists and scientists Industry partners would collaborate on joint
from other domains, focusing on priority areas research within the consortium, then take the
such as biomedical science, health research, and resulting outputs back into their own pipelines for
advanced materials, fields critical to Singapore’s competitive product development. This model
long-term innovation strategy. The equivalent pools talent, infrastructure, and funding, reducing
investment in Australia (scaled up per capita) duplication and dramatically lowering the cost of
would be around AUD $650 million. innovation, particularly for SMEs.
The AIRC would accelerate the development and
adoption of AI across the economy while building
a critical mass of AI expertise, infrastructure, and
intellectual property within Australia. It would support
faster commercialisation of AI solutions across
diverse sectors – from health to mining to agriculture
– and help strengthen our sovereign AI capability.
Over time, the AIRC would become a national asset,
embedding long-term capacity for innovation,
collaboration and economic transformation.
Establishing the AIRC will require genuine
co‑investment and active participation from
industry and academia, with the door open for
government playing a role down the road.

66 European Molecular Biology Laboratory - European Bioinformatics Institute 2025, AlphaFold Protein Structure Database ([Link]
67 imec, About imec ([Link]

Business Council of Australia 35

 06 Australia’s AI agenda

Action 14  ix R&D settings to support

F Action 15  reate the AI Commercialisation
C
AI development and adoption Accelerator (AICA)
Owner: Government Owner: Industry, academia and government
Timeframe: Complete by June 2027 (Year 2) Timeframe: Complete by June 2027 (Year 2)

Australia should elevate R&D as a national strategic Australia should establish an AI Commercialisation
priority and agree a target of 3 per cent GDP Accelerator (AICA) to bridge the gap between
spend on R&D by 2035. This should be research and deployment in priority sectors.
accompanied by a commitment to a 10-year Focused on applied development and early‑stage
national R&D strategy to drive innovation, boost scaling, the Accelerator would support
productivity, and position Australia as a global high‑potential AI projects emerging from
leader in R&D. universities, startups and research institutes,
particularly in areas like clean energy, advanced
Australia’s R&D Tax Incentive (RDTI) should be
manufacturing, digital health and agriculture.
modernised to reflect how AI is developed,
deployed and scaled in industry. The current tax Modelled on the UK’s Catapult Network68 and
definition is narrowly focused on experimental, aligned with a national R&D strategy, AICA would
hypothesis-based research and often excludes offer shared infrastructure, technical capability and
legitimate innovation, particularly in software, commercialisation support to translate research
systems integration and iterative AI model into market-ready solutions. It would complement
development. As a result, key AI-related activities the BCA’s recommendation to the Australian
such as training models, refining algorithms, Government’s 2024-2025 Strategic Examination
or integrating machine learning into business of Research and Development to establish a
processes may fall outside the incentive’s scope. nationally coordinated network of industry-led R&D
centres, providing a targeted vehicle to scale AI
To unlock AI’s full economic potential, the RDTI
innovation and ensuring promising ventures can
must evolve to recognise contemporary innovation
grow and stay in Australia.
methods and technology development pathways.
This includes updating the tax definition to align
with the OECD Frascati framework67 and reviewing
eligibility for AI-related compute, cloud infrastructure
and collaborative projects. This would ensure
the RDTI better supports business investment
in AI and positions Australia as a competitive
destination for technology‑led innovation.

67 Organisation for Economic Cooperation and Development 2015, Frascati Manual 2015: Guidelines for Collecting and Reporting Data on Research and Experimental
Development ([Link]
68 Catapult Network 2025, Catapult Network ([Link]

36 Accelerating Australia’s AI Agenda

 06 Australia’s AI agenda

Action 16  stablish a national

E
AI missions program
Owner: Industry, academia and government
Timeframe: Complete by June 2027 (Year 2)

Australia should establish a national AI missions

program to focus R&D investment on solving
high-impact national challenges with artificial
intelligence. These missions – co-designed by
government, researchers and industry – would
drive applied research and commercialisation.
Each mission should be time-bound,
outcome‑driven and closely aligned with
Australia’s broader national R&D strategy.
Modelled on international programs like Japan’s
Moonshot R&D initiative69, the AI missions program
would consolidate and target funding to build
sovereign capability, attract talent and investment,
and translate research into national advantage.
It would work closely with the NAIC, AIRC and
AICA to provide clear pathways for aligning public
and private investment, scaling collaborative
innovation, and positioning Australia as a leader in
responsible, high-impact AI development.

Contents >

69 Japan Cabinet Office 2025, Moonshot Research and Development Program ([Link]

Business Council of Australia 37

38 Accelerating Australia’s AI Agenda
 07

Conclusion

Australia’s embrace of AI presents an The foundational investments we make today are

immediate, transformative opportunity, more than short-term gains. They are the bedrock
crucial for addressing our national upon which our long-term economic prosperity
productivity challenge. While the call to and global competitiveness will be built. What
action outlines a clear vision for 2028, it’s we do now – how we develop, deploy and
imperative to recognise that our current regulate AI – will determine whether Australia
decisions will cast a far longer shadow, gains an enduring competitive edge or faces
shaping Australia’s trajectory for the next two decades of playing catch-up, directly impacting
decades and beyond. the opportunities available to future generations.

Contents >

Business Council of Australia 39

BUSINESS COUNCIL OF AUSTRALIA
GPO Box 1472,
Melbourne 3001
T 03 8664 2664
[Link]

© Copyright June 2025 Business Council of Australia, ABN 75 008 483 216.

All rights reserved. No part of this publication may be reproduced or used in any way without acknowledgement
to the Business Council of Australia.

The Business Council of Australia has taken reasonable care in publishing the information contained in this
publication but does not guarantee that the information is complete, accurate or current. In particular, the BCA is
not responsible for the accuracy of information that has been provided by other parties. The information in this
publication is not intended to be used as the basis for making any investment decision and must not be relied
upon as investment advice. To the maximum extent permitted by law, the BCA disclaims all liability (including
liability in negligence) to any person arising out of use or reliance on the information contained in this publication
including for loss or damage which you or anyone else might suffer as a result of that use or reliance.

238_AI Report_20250530_WEB