- 14 minutes ago
Malaysia–South Africa AI collaboration in predictive risk auditing: governance, financial leakage mitigation, ethical compliance, and deployable use cases.
Category
🗞
NewsTranscript
00:00Malaysia losses an estimated RM5 to RM6 billion annually to procurement leakages and weak financial controls, according to Auditor General reports.
00:10At the same time, global AI adoption in auditing is accelerating.
00:13PwC predicts AI-driven risk analytics could cut fraud losses by up to 40% within five years.
00:19Malaysia has launched frameworks like AI Governance for Education Aid and also Malaysia National AI Office in I.O., but real-world deployment remains limited.
00:27Meanwhile, South Africa is emerging as a hub for AI innovation in compliance and predictive risk.
00:33So to unpack on this matter today, we have with us in the studio, Claude Chatty, the CEO and also founder of IAT, the world developer of the world's first predictive risk AI.
00:43How are you? Good morning.
00:44Good morning, and thank you so much for having me.
00:46Thank you for coming all the way here, being in the studio this morning.
00:49But to start, we want to compare on the AI landscape between Malaysia and also South Africa.
00:53We see that Malaysia has a strong policy frameworks, but very lack on the deployment.
00:58So while South Africa, we see that they're gaining, you know, very large when it comes to attraction, when it comes to predictive risk AI.
01:07So how could you compare the two ecosystems in Malaysia and also South Africa?
01:11And what is the biggest gaps right now?
01:14Yep. Thank you so much.
01:15So my fundamental belief is that there's big parties out there that are busy developing the infrastructure for AI usage and adoption.
01:26The one key differentiator that I've noticed in my travels thus far is that from a Malaysia perspective,
01:32I think that we're still having very much the infrastructure and governance conversations, which are extremely important.
01:39But the adoption conversation is the one that needs to be on the table.
01:42So as part of my trip and my meeting with various stakeholders, including people from government as well as educational institutions,
01:51my message has been very clear.
01:54I believe that the tracks have been built.
01:57Let's ride it.
01:58And from a South African perspective, I think we've been quick to get onto that track, onto that highway.
02:04And Malaysia has definitely got the capability and the skills.
02:08Now we just need to adopt faster.
02:11So adoption has been a bit of a challenge from what I've seen.
02:17And before we start this conversation, we had this, you know, you mentioned about the Audrey, the first world AI editor by IAT itself.
02:25So we can talk more about that.
02:27And your company specifically focuses on the predictive risk AI.
02:33So what makes it different from the traditional risk analytics?
02:35And at the same time, how do you help Malaysia strengthen governance and also reduce financial leakage at the same time in governance?
02:43Fantastic.
02:43So firstly, I'm not an auditor by profession, but I've interacted with auditors and I have lots of friends and family that have been auditors.
02:54But being in business, the one bugbear for me has been that auditing has always been something reactive.
03:01You get an audit report after maybe some time after financial year end, and you expect it to mitigate issues based on that report.
03:11So what fundamentally changes with Audrey and AI, we do reporting analysis and predictive risk assessments in the now, for now, so that we can deal with issues proactively.
03:23So when you speak about revenue leakage, five to six billion riggets, we've had the privilege of dealing with the Ministry of Finance and other institutions.
03:30And we've shown them how through proactive risk assessments, we can pick up risks as and when they occur.
03:37So that enables government organizations to deal with these matters as and when they occur, rather than report on them as a loss.
03:47So I think that's a big fundamental change.
03:49The other big change is with the advent of new technologies such as AI, machine learning, this enables you to use large data sets to inform better decision making.
03:58The audit profession, once again, has had this set of rules and methodologies that's probably been around for like 50 years.
04:06But guess what?
04:07The world has changed fundamentally in the last 50 years.
04:10We can do so much more.
04:12So what we've done is we've said, okay, instead of using sample data, which is what auditors use,
04:19they use sets of sample data as part of their fieldwork and testing to test against the mitigations and the risks.
04:26What happens if we use technology to look at all the data available?
04:30Not only will we be able to be more accurate, but also we probably would acknowledge we can get things done that much faster.
04:38So the two big things that Audrey does is, one, it makes sure that you look at your books and your risks continuously for insight and also predictably for foresight.
04:51So you can see things potentially before they can happen.
04:54And the second thing is we look at all the available data.
04:58So we're not being big brother or the police, but what we are trying to do is build a more transparent and more productive organisation and world that we live in.
05:08But before we go deep further on the mitigation strategies, what are the stats and data based on reports on the procurement leakages and also weak financial controls that you see?
05:19Look, I mean, we've obviously done this under strict guidance and conditions.
05:24But what I can tell you is I can speak about the themes.
05:28So some of the themes that we we've picked up and shared with the ministry revolves around how we can do things faster, how we can leverage our Malaysian people more as part of our supply chain.
05:45And of course, how we can ensure that there's better productivity, meaning every rigged spent is a rigged well spent and ideally better spent within the country.
05:57So those are three big things that I think that I can call out specifically on the show without being hurled before the colds.
06:05Well, we see that a recent Prime Minister Anwar's visit to Africa, which signals deeper tech and also trade ties.
06:14So in your opinion, what we did opportunities that you see for Malaysia and also South Africa, AI partnerships arising from this diplomatic momentum.
06:22Could this diplomatic momentum lead to a Malaysia-Africa, you know, AI corridor and especially when Malaysia has an ambitious national goal of becoming AI nation by 2030 and also a regional AI hub in Southeast Asia?
06:34So firstly, I think I think I can speak on behalf of all my countrymen in that we've got a deep and decorated history with the Malaysian people.
06:46As a South African, we're very proud of the ties we have with Malaysia.
06:50I learned that our first ballot papers were printed in Malaysia.
06:54That's post apartheid.
06:56So and our policies have been modeled on many of Malaysia's successful policies.
07:00From an AI perspective, and specifically with Audrey, our Audit Nagara or Auditor General and the Malaysian Auditor General, I think through this process, having dealt with both, we're able to connect both.
07:17But more so from a technology and knowledge transfer perspective, we signed a collaborative agreement with MIT, your government agency earlier this year.
07:27And part of that entails a skills transfer and knowledge transfer.
07:31Not only that, we're going to the national AI organization a bit later today to also confirm our commitment to train Malaysian people.
07:40In fact, we're taking our first two students along to the NAIO to to show them that, well, this is more than just a product push.
07:48This is a partnership push where we've been able to harness and make use of the available technologies.
07:54We've not built the infrastructure, as I mentioned.
07:57We are driving on that highway at breakneck speed.
08:00And I think we want to take our Malaysian people and young Malaysians on that part with us.
08:07And so this is this is the start of that journey.
08:10Since you've highlighted on the talent development, what concrete steps are you to take to actually build Malaysia AI capability rather than just exporting technology?
08:19Exactly.
08:20So, I mean, it's a multi-pronged approach.
08:22So, firstly, with the partnership with MIT, our agreement at a very high level is absolutely predicated upon setting up offices here and setting up a skills development center of sorts.
08:38So we're meeting with MIT tomorrow to firm that relationship up.
08:43So we're busy setting that up.
08:46So that means that for now, yes, we're going to have our South African AI engineers in country for the shortest time.
08:54But through our partnership agreement also with a Malaysian company called PCG, we're doing a train the trainer exercise.
09:01So they're going to be helping us train future AI engineers and future technologists to drive this.
09:09You see, my mantra and yesterday we were privileged to be at the Serum International Conference.
09:15And my mantra to everybody was around, look, let us not get too immersed in the architecture of the technology, but let's rather immerse ourselves on how this technology can be used now for good, for sustainability, and to increase productivity.
09:34There's people that are naysayers around AI.
09:38Is AI going to take over our jobs?
09:39Is it going to be the age of the machines?
09:45But what I honestly believe is that it is the age of the human and machine collaboration.
09:50And the AI that we are using now, well, it's generative AI and we're moving to agentic and hopefully multi-agentic AI, which I could talk to you about for a very long time.
10:00But that's all about elevating us as human beings.
10:05It's about elevating a professional to augment what you do and make sure that it makes you a better human and a better professional specifically.
10:15So it's about accelerating that to our youth and making sure that let us not get too involved in the back-end stuff.
10:23Let's see how we can use this now for business, for education, for healthcare, and for our economy.
10:32But that's the thing about AI.
10:34The dynamic of AI landscape is super complex, if I must say.
10:38So when it comes to AI compliance and auditing, it involves very super high stakes.
10:44So how do you ensure principles like perhaps human oversight, also fairness and transparency that's embedded in your solutions, perhaps in Audrey?
10:52I mean, look, fundamentally, we're an audit platform.
10:58So by nature, the designers and the back-end have been designed by auditors.
11:04So governance is extremely important.
11:08Now, whilst I think kind of standards authorities and the governments are regulating, what's very important, I believe, is self-regulation.
11:17So with AI, you build guardrails around what AI can and can't do.
11:23But also, I think it's important to adhere to national and international laws about where to draw the line.
11:31So as auditors, what we do is that we also have built-in structures within our AI.
11:37For example, traceability audits.
11:39So there's something that's called AI hallucination.
11:43So as we were speaking to people yesterday, for example, if you use five different AI and GPT platforms, for example, they may give you five different responses.
11:56But where Audrey comes in play is that our traceability audits will tell you.
12:00If someone said A, B, and C about Nabila or Claude, we can trace that.
12:05We can trace that to the source because I think ultimately, and this is where I'm a very strong proponent, AI does not replace the human being.
12:15The AI augments the human being.
12:18The human being still needs to be empowered to make the final decision, as is the AI auditor.
12:24It is not an auditor.
12:26It's an audit and assurance assistant.
12:28An auditor still needs to sign it off.
12:30An auditor still needs to verify that.
12:32And I think this is the happy place when I speak about human and machine collaboration, is that the human being still holds the strategic insight.
12:42We are the real intelligence.
12:44And the artificial intelligence, we allow that to do the heavy lifting, the hard work, so that we can do the smart work,
12:51which is the thinking, the decision-making, the verification, and to clear all the blurry spots.
12:58So Audrey, is it accessible to everyone?
13:01Obviously, it's a subscription-based model, but it is accessible to everyone.
13:05So the way we've built it is that I've been a strong proponent of the underdog.
13:10This is how this was built.
13:12And it was built on a dream that currently there are a few big consulting and audit houses.
13:18I think we know the big four, the big five.
13:20And the dream started with was, what if we made a 10-man or a five-man audit shop behave like a 100-man audit shop
13:28and be able to execute like a 100-man audit shop to play with the big boys and to compete with the big boys?
13:35And how could we use technology to do that?
13:37And so through this iterative kind of process, we realized, hang on, technology can help us.
13:45So it can help a small business, a micro-business, all the way to a multinational business, depending on the application and depending on the needs.
13:54So very cool.
13:56One very other cool thing, and this is not just with Audrey, with other technologies as well, is the rise of agentic AI,
14:03where, for example, we are now able to, if you're a subject matter expert, we give you a persona or some other people call it a co-pilot.
14:13That co-pilot, once you open your system and start interacting, it actually starts learning how you do things.
14:20So that one day, if you may change jobs, if you decide to retire, or for whatever reason, the AI co-pilot or agent continues to work.
14:33So that's just the power of technology.
14:36Amazing. Thank you so much, Claude.
14:38Thank you so much.
Recommended
1:19
|
Up next
1:35:58
1:30
49:30
26:10
1:07
1:22
1:36
1:11
5:47
Be the first to comment