Key points
- ASIC won't hold back customer-centric AI innovation, but cutting-edge technology can’t leave your customers bleeding.
- We all have a duty to harness the technological advances we are seeing - and use them for community and customer benefit.
- We are not rushing to more AI regulation but acknowledge more regulation may be needed in future. In the meantime, regulators will need to be bolder and more imaginative with how we use our existing powers.
Thanks for that introduction. Before I start my formal remarks, I would like to begin by acknowledging the Traditional Owners, the Gadigal people of the Eora Nation, and their ongoing connection to and custodianship of the lands on which we meet today, and to pay my respects to elders past and present. I extend that respect to Aboriginal and Torres Strait Islander persons present.
Of course, the other important acknowledgement is Anna Bligh [outgoing Australian Banking Association Chief Executive Officer]. Anna, you’re going to have to be putting up with this for the next two days. But seriously, you’ve been a committed and constructive leader of the ABA, particularly during a transformational period for the banking sector. We won’t go over the history again but I am particularly grateful on behalf of ASIC for the collaborative way you have engaged with us on so many fronts. I won’t try to summarise them but including most recently on one of my pet projects, which of course is regulatory simplification, otherwise known as deregulation. You’ve gone to three meetings, you’ve done some great work for us there and I look forward to continuing that work with the ABA under the leadership of Simon Birmingham. Simon - welcome. I look forward to working with you as well.
Introduction
‘AI has been around now for 75 years or so. In the first few decades, breakthroughs came every 5-10 years. Now they seem to be emerging every 6 to 12 months, some as technological innovations, others as dramatic new applications. […] My view is that we will be basking in an AI summer for the foreseeable future.’[1]
So writes Richard Susskind in his recent book on How To Think About AI: A guide for the perplexed. I’ve read this book and I have to tell you, I’m still perplexed. However, we would both go further.
I think it’s fair to say that AI represents one of the most momentous and significant areas of change in today’s world. Like any tool, of course, it can be used well or badly. The opportunities presented by AI, particularly when it comes to customer service, are undeniable. But with every new opportunity comes new risk, and so AI is also an opportunity to reflect on how you can keep serving your customers.
In short, I believe AI is both a catalyst for a new era in customer-centric banking – and a reminder that customer-centric is exactly what banking ought to be about. It’s both sides of this coin that I’ll be addressing today.
AI – its perception and opportunities
First, let’s consider AI as we know it today. By that I don’t mean the state or progress of its development – that would be a risky undertaking which would likely become obsolete before my remarks have even ended! Instead, I mean to look at how AI is perceived by everyday Australians, and the opportunities it presents for better banking.
I start with perception, because perception precedes adoption of technology. When Christopher Sholes and his associates invented the typewriter in the 1860s, they struggled to get financial backing for it and almost went bankrupt. Why? Because the perception was that there was no need for it – what would all those stenographers do with all their time? A similar story is told about the invention of the sewing machine.
The point is that without the trust of the customer, technological developments can struggle to make progress. You could have the best technology in the world, but if people don’t trust it, they’re not going to use it.
And trust is what seems to be missing. The latest research by RepTrak and KPMG shows that over the last 12 months, Australians’ trust in AI has declined.[2] Australians have lower levels of AI use, training, and confidence than their peers around the world. Five out of 10 Australians use AI – but fewer than four out of 10 (36%) trust it.[3] The number of Australians who think that AI doesn’t improve customer service or the way businesses are run has also risen since last year.[4]
This is a problem, because we’re hearing a lot of talk right now about AI as the ‘next big lever for economic growth’. Much has been said recently about Australia’s ‘productivity problem’. Our labour productivity growth has halved in the past two decades[5]. Low productivity is not an abstract number on a page. It’s people working more but feeling they’re getting less. It’s everything getting more expensive but not better.
The promise of AI is that it can raise living standards, and ‘deliver a significant productivity boost’ with ‘higher real wages and more opportunities for more Australians.’[6] But this can only happen when Australian workers embrace it - rather than feel that AI is being imposed on them.[7]
So the perception is a bit of a mixed bag. On the one hand, we have business leaders across a range of industries singing AI’s praises, opportunities, and potential – while on the other hand, the number of Australians who distrust the technology appears to be growing.
This might sound like a problem – but I would suggest that, in financial services, and in the banking sector in particular, it is instead an opportunity for industry to shift community perception.
Banks are some of the most sophisticated operators in Australian business – and if anyone has the systems, technology, and processes to lead Australia’s AI revolution, it’s all of you. You have been using machine-learning for years and already have some of the necessary governance structures in place. You have teams of data scientists, and the infrastructure to support technological advances. And, since Australians interact with their banks in some form or other every day – if you get AI right, that will go a long way toward addressing Australians’ general hesitation around AI.
The good news is, we’re already seeing a boldness in the banking sector to move beyond the basics. A cursory flick through the news shows how AI is already being used to solve genuine customer problems in our biggest banks.
Look at Westpac piloting AI to help its scam and fraud support team process and analyse information provided by customers about potential incidents, by providing live transcripts, or alerts when key indicators are detected.[8]
Or NAB using AI to find ‘needles in a haystack’ and assessing customer complaints for issues that could become systemic problems.[9]
Or CBA beating scammers at their own game by unleashing an army of ‘bots’ who sound like real people – Australian accents and all – to tie them up on lengthy phone calls, thereby preventing them from contacting real people[10].
Here, I want to acknowledge for a moment Macquarie University’s Cyber Security Hub Executive Director Professor Dali Kaafar, who first conceived of and developed this technology. It is wonderful to see Dr Kafaar’s genius and persistence come to fruition in a way that will no doubt help many, many Australian consumers.
Do we need more regulation?
So, all this is very promising – and very exciting. And as the theme of this conference highlights, it’s also likely to be disruptive in terms of how banking is done.
Which leads me to the question people often ask me: Do we need more regulation around AI, to ensure it is adopted safely and wisely?
My position – and ASIC’s – is that we should not rush into more AI regulation.
First, because our existing laws are technology-neutral. In essence, some guard rails already exist – and ASIC will enforce them as they stand.
An important example is the obligations of directors under the Corporations Act, which apply broadly – and as companies increasingly deploy AI, that use of AI is one of a range of issues that directors must pay attention to, as part of their existing duties.
Second, the reality is, as I have said publicly before, the more specific and narrowly targeted regulation becomes, the more complex it becomes. It gets harder to understand, harder to comply with, and harder for regulators to enforce.
This becomes a burden borne by everyone – legislators, policymakers, businesses big and small, regulators, and consumers and investors. It’s played out across the economy in lost time and productivity. The mosaic of complexity also impacts what matters to consumers – including how we combat scams, predatory lending, and unfair contracts. And it becomes a handbrake on innovation, when you have got to spend more on compliance lawyers than coders.
So, while we're not rushing to new regulation, that’s not to say that new regulation won’t be needed.
To borrow again from Richard Susskind, ‘one clear practical issue is that we simply cannot roll out our entire policy making apparatus every time there is a major new advance in AI. As the new developments come at us at an ever-increasing pace, we will have neither the time nor the capacity’.[11]
That’s the challenge. We have yet to create the regulatory architecture, language, or strategy that will enable us to regulate confidently and in real time a group of technologies that is evolving as rapidly as we’re seeing.
In the meantime, regulators will need to be bolder and more imaginative about how we use our existing powers. You can count on ASIC on that. In other words, we’ve got to kick the tyres a bit and see just how far our existing technology-neutral framework will flex.
It’s already apparent that AI could present some challenges for regulators in data and information gathering. That could, in theory, stretch and test our ability to identify and assess systemic harms. These are bridges we are yet to cross but they are certainly looming.
Ultimately, we don’t want regulation to have a chilling effect on AI innovation and the potential benefits that it would have for all of us.
Will you get in trouble?
So, while we are grappling with the question of what new regulation might need to look like, you might ask: ‘Will we get in trouble in the meantime?’
It’s a legitimate concern, given that ASIC’s review last year of 23 AFS and credit licensees that were using or planning to use AI found a number of gaps in how they assessed risks to consumers from AI. For some licensees, their governance arrangements lagged well behind their AI use.[12]
ASIC is continuing to monitor how entities are using AI, and we will take enforcement action where appropriate, where we see misconduct occurring. So yes, you will get in trouble when technology is not used responsibly. Which then leads us to the question, ‘What’s the benchmark for good AI innovation?’
And I think the answer is pretty clear: good AI innovation keeps the customer front and centre.
What does this look like? In essence, it involves using AI to better serve the customer.
ASIC will soon be releasing the follow-up to our Better Banking for Indigenous Consumers report, which will address the impact of fee harm to low-income customers nationwide. Millions of dollars of bank fees are being refunded to customers as a result of this work, with a significant number of eligible people being migrated into low-fee accounts.
It would be a true win for customer-centric banking if your algorithm prompted you to serve your customers better before an ASIC report did.
If it had, maybe people like Cassie – not her real name – a single mother from South Australia, would not have been charged more than $2,600 in overdraw fees over five years. To put that into perspective, that’s equivalent to 13 and a half 8-hour days at the minimum wage.
Or Charlotte (again not her real name), a jobseeker from Western Australia, might not have accrued over $4,000 in dishonour fees over the same period.
It’s great that these refunds are being made, and that the systemic issues identified in our Better Banking report are being addressed. [13] But isn’t it disappointing that people like Cassie and Charlotte had to wait for that report to see a positive change in their lives?
So, my question to everyone in this room is this - what are you doing to ensure people like Cassie and Charlotte are better protected and supported in the future? Is AI going to be part of the answer?
To put it bluntly: cutting-edge technology can’t leave your customers bleeding. We’ve already seen that customer trust in AI and its potential to improve customer service is eroding. If banks get this wrong, we’re likely to see a significant setback in AI legitimacy and trust.
Talking up the benefits of AI must be accompanied by showing customers how it makes their lives better. And more than that - AI innovation has to actually make your customers’ lives better and not just be a marketing or data-gathering tool. And that requires careful attention to ethics when introducing these systems to ensure customers actually benefit over the long term.
Many of the innovations I mentioned earlier are good examples of AI systems for consumer benefit. We all have a duty to harness the technological advances we are seeing - and use them for the benefit of the community and customers.
Conclusion
In conclusion, AI is an opportunity to reinforce your focus on customer-centric banking.
If done well, AI adoption and digital transformation can be a blueprint for better banking. ASIC won’t hold back customer-centric AI innovation.
But, don’t be complacent and don’t forget - the customer must be at the centre of everything that you do – and ASIC will be watching.
[1]Susskind, R (2025). How To Think About AI: A guide for the perplexed. Oxford University Press. (p.21)
[2] Data refers to June – August 2024 Source: The RepTrak Company; KPMG, Global study reveals Australia lags in trust of AI despite growing use, https://kpmg.com/au/en/home/media/press-releases/2025/04/global-study-reveals-australia-lags-in-trust-of-ai-despite-growing-use.html
[5] Productivity | Australian Bureau of Statistics
[6] Cf. Business Council unveils Australian AI report to secure leadership by 2028, https://www.bca.com.au/business_council_unveils_australian_ai_report_to_secure_leadership_by_2028
[7] https://www.uts.edu.au/globalassets/shared-media/documents/CSJI/essentialresearchuts_invisible_bystanders_0524.pdf
[8] Westpac pilots AI to analyse inbound call content - iTnews
[9] National Australia Bank is rolling out generative AI in these three areas
[10] Commonwealth Bank uses AI bot with Australian accent to trick phone scammers
[12] Report 798 Beware the gap: Governance arrangements in the face of AI innovation. ASIC also recently undertook a similar review looking at 40 market intermediaries. We found similar results, and similar concerns raised in relation to gaps in governance arrangements.
[13] Report 785 Better banking for Indigenous consumers
