Professor Terry Flew, Co-Director, Centre for AI, Trust and Governance, The University of Sydney
NSW Parliament House (via Zoom link), 7 October 2026.
I would like to begin by thanking the Joint Select Committee on Artificial Intelligence for inviting me to be here today. I would like to acknowledge that I am on the ancestral lands of the Gadigal people. I would also like to commend the Committee for the timeliness of this consultation.
My name is Terry Flew. I am Professor of Digital Communications and Culture at the University of Sydney. I am Co-Director of the Centre for AI, Trust and Governance which was established through the Faculty of Arts and Social Sciences in 2025. My Centre Co-Director is Professor Kimberlee Weatherall from the Sydney Law School. I am also an Australian Research Council Laureate Fellow, and my 5-year project is on Mediated Trust: Ideas, Interests, Institutions and Futures.
The Centre for AI, Trust and Governance at the University of Sydney is dedicated to ensuring that artificial intelligence is developed, deployed, and governed in ways that advance public value, strengthen trust, and support Australia’s economic, social, and democratic priorities.
It is an interdisciplinary centre which recognises that as AI rapidly transforms industries, institutions and everyday life, there is a need for independent research and for public platforms that enable knowledge exchange, policy engagement, and cross-sector collaboration focused on the human, organisational, and societal dimensions of AI.
The Centre’s ultimate purpose is to help shape an AI enabled future in which technological progress is matched by strong governance, public trust, and meaningful societal benefit.
The pace of change
It is sometimes said that the time to act on artificial intelligence was yesterday. I prepare a lecture on this topic for students at the University of Sydney. In the slides that I prepared for those students in 2023, it was estimated that by 2031 artificial intelligence would be worth $US1 trillion. When I revised those slides in 2025, it estimated that the artificial intelligence market would be $US3.3 trillion dollars by 2027.
With my colleague Dr Rob Nichols, we prepared the submission for the Joint Select Committee addressing all of the Terms of Reference outlined. Recognising that there are other inquiries taking place into such matters as AI infrastructure and data centres, I want to focus my comments today on a small number of issues. Those issues will be: trust; the AI dividend; copyright and IP futures; and the distribution of opportunities and benefits.
In his address at the University of Sydney outlining the Federal Government’s AI strategy, Prime Minister Anthony Albanese focused on three issues in particular: copyright, data centres, and the future of jobs and work. There were also related issues under consideration which included national security, AI sovereignty, critical infrastructure, the rate of AI adoption and associated questions of risk, and the capacity of middle powers to frame what might be referred to as a third way in AI governance.
AI and trust
With regards to trust, it is a cross cutting and multifaceted concern that runs through all of the Terms of Reference, and should not be treated as narrow function of data security. Australian AI and data policy has sometimes paid insufficient attention to questions of fairness and the distribution of benefits from new technologies. This runs the risk of eroding the societal trust on which technology adoption – and ultimately our democratic system of government— depends.
Australia faces a significant trust paradox with regards to AI. In terms of attitudes towards AI in general, it has been found that Australia sits in the bottom third of countries surveyed with only 30% expressing trust in AI in general, and only 17% expressing high acceptance of the use of AI.
At the same time, a recent survey by Anthropic of use of its Claude AI system found that Australians were proportionately among the highest users in the world. Australia accounted for 2.5% of all Claude uses, and were proportionately higher users than Singapore, as compared to 0.5% of global Internet users and 0.33% of the world’s population.
Australia thus faces a significant AI trust paradox, where high levels of personal use of AI systems coexist with significant concerns about AI as a General-Purpose Technology (GPT). Australia also sits within the Asia-Pacific region, where rates of trust in AI and enthusiasm about the use of AI are among the highest in the world.
In our submission, we proposed that the Office of AI should work with independent researchers in the development of a Global AI and Digital Trust Index. Such an Index is needed in order to reliably and independently identify changes in Australians attitude to trust in AI and other digital technologies over time, and to benchmark these trends against comparable international evidence. This index would operate along the lines of the highly respected Scanlan Index of Social Cohesion that has been a resource for Australian policymakers and the wider community since 2007.
Realising the AI dividend
Whether Australia experiences rising or declining levels of trust in AI, and indeed trust in social institutions more generally, will critically depend upon the equitable distribution of benefits of AI, or what we might term the AI dividend. There is considerable opportunity for local communities to extract benefits from investment in AI infrastructure as part of the social licence that is expected from these large international companies to operate on Australian shores. Investment in jobs in communities and in culture will be particularly important.
In this regard we want to commend the importance of investing in First Nations communities and in learning from the insights that First Nations communities have generated over millennia around caring for country.
Any lasting approach in the copyright space must also take account of the unique nature of Indigenous Cultural and Intellectual Property and will need to be aligned with the United Nations Declaration on the Rights of Indigenous People (UNDRIP) to which Australia is a signatory. I particularly commend Article 31 of UNDRIP to this Committee.
AI and copyright
A new copyright and intellectual property bargain is urgently required for the AI era. There has unquestionably been theft of the work of creative people by companies that have put their work into Large Learning Models without permission or recompense.
At the same time, the Copyright Act 1968 has long sat uncomfortably with the affordances of digital technologies and is particularly challenging to apply in the AI era. One of the reasons for this is that the provenance of content, or the extent to which it can be clearly traced back to an original source, becomes more and more challenging. We can note, for instance, the declining use of links from search engines as a source of traffic to news sites, as news is increasingly sources from AI systems themselves in the form of AI-generated summaries of available journalism.
This points to the need for new forms of licencing, and the empowerment of creatives to have bargaining rights in dealing with the tech companies. There is also a leadership role for government in establishing protocols for the proper acknowledgment and recompense of creatives for the use of their work. There are emergent models starting to appear which I can discuss with the Committee.
Equitable distribution of AI benefits
Ensuring that there is not only widespread AI adoption, but active engagement with AI among the Australian populace, will be critically linked to the distribution of opportunities and benefits. This is partly about skilling up the populace, not just in technical skills, but also in understanding the ethical. moral and epistemic challenges associated with these profoundly transformative technologies. Skills development for AI needs to incorporate sovereign evaluation capability as well as skills and capabilities in AI safety, trust and ethics.
The South Korean AI for All initiative provides a very interesting and important example for Australian policymakers to consider. AI for All provides one example of an equitable and inclusive national AI strategy distributes the benefits of the new technology equitably and is also cognizant of sovereign capability. Put simply, the South Korean AI for All strategy provides each Korean citizen with their own AI agent drawing upon locally sourced LLMs.
The public sector could provide an important test bed for strategies around what actually drives durable adoption and equitably distributed productivity benefits. This could include explicability of processes and outcomes, and a focus on data quality, interoperability and workforce capability. There are opportunities for public sector agencies to act as test beds for safer AI with transparent risk assessment processes, data governance, and assessment grounded in independent research access, including the development of public sector LLMs.
AI and democracy
Finally, I would like to say something about AI and democracy. There are real risks that AI-generated misinformation, as well as what has come to be known as AI slop, will further erode public trust in institutions and degrade the wider information ecosystem. There is no single agency which can address or misinformation issues and the challenges of what are being termed epistemic risks.
One recommendation we would make is that current Australian political advertising laws be adapted to prohibit the use of AI-generated deepfakes in election campaigns. Such activities go beyond the legitimate bounds of public campaigning and prove to be genuinely threatening to the individuals and groups who are the targets of them.
The future of democracy is not only challenged by AI-generated misinformation but also concerns about a wider commitment to truth that can corrupt the epistemic bases of democratic societies. Greater regulation of AI companies is clearly important to addressing such concerns. But this remediation of communication through AI presents new challenges to our educational institutions, and knowledge, media and communications institutions, as well as those of government. It is vitally important to address such challenges as a matter of urgency so that we can have a trustworthy and vibrant society in an increasingly AI-enabled economy.
About the presenter
Terry Flew is Professor of Digital Communication and Culture, Australian Research Council (ARC) Laureate Fellow, and Co-Director of the Centre for AI, Trust and Governance at the University of Sydney. His books include Media Economics (Palgrave, 2015), Understanding Global Media (Palgrave, 2018), Regulating Platforms (Polity, 2021), SAGE Handbook of the Digital Media Economy (SAGE, 2022), Digital Platform Regulation: Global Perspectives on Internet Governance (Springer, 2022, and Valuing News: Digital Platforms and Journalism Futures (Springer, 2026). He was President of the International Communications Association (ICA) from 2019 to 2020, and is an ICA Fellow, elected in 2019. He is a Fellow of the Australian Academy of the Humanities (FAHA).
