Arie Freiberg on regulatory tools, failures, and the future of AI governance

Emeritus Professor Arie Freiberg AM is one of Australia’s most respected thinkers on regulation. With a career spanning academia, public policy, and law reform, he has shaped how generations of regulators understand their work and why it matters. 

He’s also the author of Regulation in Australia, doubtlessly the country’s most comprehensive text on regulatory theory and practice, and has served in leadership roles at Monash University and the Victorian Sentencing Advisory Council. He has also taught in the Australia and New Zealand School of Government. 

The second edition of Freiberg’s book arrives in a moment of global regulatory strain, where trust in institutions is faltering, regulatory failures routinely make headlines, and the rise of artificial intelligence is challenging long-held assumptions about decision-making, accountability, and public interest. Freiberg does not offer silver bullets, but he does offer frameworks. He maps how regulation actually works in practice and how it could work better. 

In this interview with The Modern Regulator, Freiberg reflects on how the field has evolved since 2017, how regulators can learn from failure, and why building regulatory capability is more a civic than technical exercise. 

The Modern Regulator: What has changed since the first edition of your book? 

Arie Freiberg: Well, certainly the major change has been the development of RegTech and AI. AI is transforming societies and its regulation is absolutely at the centre of discussions, whether it’s at the OECD, or in law reform commissions or policy agencies. There are dozens and dozens of harms identified from the digital economy and AI. So it’s a real challenge. 

Another difference is there was a huge focus ten years ago on behavioural economics. That was the flavour of the month. It’s now become more of a mature part of regulatory tools – a standard part of the regulatory repertoire. 

TMR: How do you see the role of technology and data evolving in regulation? 

AF: Data is the new oil, and people talk about data lakes and data lakes being poisoned and toxic. In terms of AI, the gathering, analysis, and use of data in risk-based regulation is absolutely essential. That’s now the key. 

The use of AI and data in predicting the risks and dangers that we aren’t fully aware of and then preparing for that is a major development. It’s come on very quickly. We were talking about big data maybe ten years ago, but now it’s absolutely essential for regulators. I think these are major challenges for any regulator now. 

TMR: Are regulators becoming more sophisticated in how they use regulatory tools? 

AF: Some regulators – the more traditional, sophisticated large ones – are aware of the tools. They don’t necessarily use them. They certainly don’t articulate them, and they don’t talk about them. They don’t write about them. But I’ve tried to identify this broad range of tools, what they are and how they are – or should be – used. 

The one thing that’s often forgotten is the role of evaluation because it’s too hard, or it’s not funded. I really do think a lot of people coming into regulation have no idea of the range of tools, their uses, and how they relate to the other concepts of compliance and enforcement and soft law. 

TMR: What is transactional regulation? 

AF: It’s basically a consensual or semi-consensual form of regulation. You find this in the granting of contracts by government. The regulation of those activities is done through contracts. So you would put all of the requirements in the contract, rather than legislation. You might say, “Okay, if you’re going to be doing this service, you’ve got to provide a reconciliation action plan. You’ve got to be doing this on climate change.”

So there are thousands of NGOs that provide services, and each of those has a granting deed. And those granting deeds contain public policy requirements. That’s a form of transactional regulation. It’s not the traditional rule-based command and control, but it is a very major mechanism. You can see its power by (United States President Donald) Trump’s threats to withdraw funding from universities to force them to change their behaviours.

TMR: What’s one concept you wish every regulator understood?

AF: That you need to understand the full range of regulatory tools available to you – the concepts and frameworks. There are problem-solving methodologies that you can use based on the methodologies used in regulatory impact statements. One of those is understanding the full range of regulatory tools. Another is about how you engage with your stakeholders. Another is about identifying the harms.

You need to know the harms, the actors, and the tools. If you get that little triangle or pyramid right, then you’ve got a good chance of solving the problem. I think the message there is: when someone says “solve the transport problem, solve the climate problem,” have an understanding of that methodology. 

TMR: Are we getting better at learning from regulatory failure?

AF: When I wrote the preface (to the book), I must have been in a very depressed mood. Because we are still prone to a lack of understanding of what causes the problems and how to respond to them. The temptation is to say that the regulators are all bad, and you get rid of them and start again. That’s not very helpful.

Many major regulatory failures are followed by inquiries which make numerous recommendations to both allocate blame and prevent the harm from reoccurring. Many of these are ignored because they are too expensive or too politically difficult or sensitive or face too much resistance from entrenched interests. 

There’s good literature on regulatory failures and the importance of evaluation. We need to pay more attention to the learnings about what works and what doesn’t. 

TMR: What does regulatory capacity-building mean in practice? 

AF: It’s regulatory literacy. It’s being aware that there’s a field out there. It’s not just one agency with one mission. There are thousands of regulators. It’s doing the courses that the ANZSOG runs. It’s doing the Harvard program. It’s about organisations doing internal training. 

When you’ve got people who come from medical backgrounds, social work, psychology, enforcement – how do they all speak the same language? The only common language is regulation. That’s the challenge. You can’t expect people to be trained as regulators in school. We need training in regulation – but we’re not there yet. 

TMR: What emerging risks are we underestimating? 

Crypto. I don’t understand it but many do and it’s becoming more popular and entrenched. The US is backing off on crypto regulation. And a lot of the libertarian jurisdictions think it’s great because it avoids government. There are many pitfalls ahead and there are many people who will lose money and a few that will make it. 

TMR: What gives you optimism about the regulatory profession? 

AF: That there are about 7,000 people in the ANZSOG community of practice in Australia. That there are courses on regulation  being developed and taught over the world. My bibliography barely touches the surface of what’s available. It shows there’s an incredible amount of research. 

So yes, I’m optimistic. The people who go into regulation are committed. They want to do the right thing. And I think there’s a growing maturity in the field. That gives me hope. 

To explore these ideas in greater depth, pick up a copy of the second edition of Regulation in Australia at Federation Press.

Subscribe to The Modern Regulator for more interviews, insights, and commentary on the future of public interest regulation.