• 4 months ago
Gullnaz BAIG, Executive Director, Angsana Council Evi FUELLE, Director, Global Policy, Credo AI Zee Kin YEONG, CEO, Singapore Academy of Law Moderator: Clay CHANDLER, FORTUNE
Transcript
00:00So, this is a very rich and complicated topic, just like the region itself.
00:06I want to start, maybe, if we could, just thinking about the opportunity.
00:10We heard earlier in the discussion about whether customized large language models could be
00:16devised for Southeast Asia, that there were all kinds of commercial opportunities that
00:21would fall in line with that.
00:23And I'm wondering if the same logic applies to governance models.
00:29Do any of you have thoughts as to what the advantages might be of being able to come
00:34up with some kind of common framework?
00:37Maybe we can think about what that framework might be down the road a bit, but just talk
00:42about why the region might have an incentive, whether it's commercial or otherwise, to kind
00:47of band together.
00:48Vilnes, do you want to take the first shot at that?
00:51Yeah, sure.
00:52Firstly, thank you for having me on this panel.
00:55I sit at a unique intersection, being the Director for Ansana Council.
01:00We sit pretty close to start-ups, various tech companies, being from the tech sector
01:07ourselves, but at the same time, we engage quite regularly with various governments in
01:12the region.
01:13And I think what we see is, in terms of ASEAN integrity and there being a digital economy
01:19per se, by 2030, we could be the third largest economy in the world if we were one country.
01:26There is extensive business potential, but the challenges lie in the fact that it's actually
01:32very hard to go across border.
01:35So the ability for us to leverage on the true potential that is offered by Southeast Asia
01:40sometimes hampered by lack of, say, for example, simple things like business recognition across
01:47border, some issues around cross-border data flow, cyber security standards.
01:53Just to introduce a few that I'm quite sure my co-panelists will dig into.
01:59So if we can get a framework together, there is this huge market out there for us.
02:05And you do see some bigger markets in the region leaning towards protecting their own
02:12domestic market by putting up regulatory barriers.
02:16So that sometimes gets in the way of developing a regional framework.
02:20Really interesting.
02:21And it's an interesting, it's a big opportunity as a market, but it's also a big opportunity
02:27from an investment standpoint.
02:28Like if you're a VC investor, the ability to be able to invest in a company that operates
02:34across an entire region versus just in a single market is huge.
02:39And I like your point about it being the third largest economy.
02:43It's not only that it's large if you think about it in an aggregate.
02:47It's also that it's got great demographics, most of the individual economies.
02:51And nearly all of them are right now growing at very respectable high growth rates compared
02:55to many of the other world's economies.
02:57So if you think about that down the road, it really, it kind of adds up.
03:02What's your take on it?
03:03Yeah.
03:04If I could add to the comments you've already made, I think at Credo AI, we're working with
03:09enterprises that are highly regulated and they are global in scope and they're operating
03:13in these markets all over the world.
03:15So not only in this region, but globally, they're looking for that commonality.
03:19How can they create a governance framework that is going to stand the test of time, that
03:23is not a brittle framework that only works in one jurisdiction or only satisfies particular
03:28risk requirements.
03:30So we're constantly looking for those elements of commonality.
03:33And I think that the benefit of that for businesses is building trust over time in different markets
03:38with different regulators, but also having certainty with which they can operate and
03:43continue to operate in the long term.
03:45So I think we've seen a lot of commonality even between things like the EU AI Act and
03:50the White House Executive Order on AI in terms of what the requirements are from enterprises.
03:56Things like having comprehensive AI governance, risk management systems, transparency reporting,
04:01some form of public notice, impact assessments, an AI registry, things like this are sort
04:07of the common elements we're already seeing emerge.
04:09Zeking, how about you?
04:11I know you've looked carefully at this.
04:13Yeah.
04:14So I personally have been working in this space of convergence for many years around
04:19data policies and AI.
04:21And the Academy of Law actually has a think tank, the Asian Business Law Institute, that
04:28does work firmly in this space.
04:30I think it is utterly important for us to really take a step back and look at Southeast
04:33Asia as a whole, right?
04:36We need to understand the maturity of the economies and also the technology, right?
04:43And I think the majority of Southeast Asian nations are likely to be deployers and users
04:50of AI systems.
04:52Now the good thing is that there is a fairly consistent set of principles around responsible
04:59AI.
05:00You can point to, for example, the OECD AI principles as a set of principles.
05:07Now how do you take a set of principles like this, bring it into Southeast Asia, and achieve
05:13two goals?
05:14One, to make it a lot easier for companies that are trying to bring products and services
05:19into Southeast Asia to very quickly localize the same products and make it a lot easier.
05:25The second is, of course, how to assist companies in Southeast Asia to also adapt to that and
05:31grow out and expand beyond Southeast Asia.
05:36We have actually a very, very useful platform, ASEAN, right?
05:40And through ASEAN, it is possible to develop a governance framework for responsible AI
05:48modeled after and looking very much like the OECD principles and use that as a mapping
05:54standard so that as individual companies develop their own national frameworks, they're able
06:00to map into something like that.
06:03The ASEAN framework can be moderated, right, to suit the situation for ASEAN, which is
06:08largely deployments, right, of AI system, and also to adjust it such that it's a lot
06:17easier for the companies and the governments working in this space.
06:22Really interesting.
06:23I think there are several things there that sort of caught my ear.
06:29One is, you know, ASEAN as a framework for this, there have been, there's an ongoing
06:36discussion, really, about how ASEAN might kind of work together to come up with a common
06:40framework.
06:43How far along do you think that is, and is it making progress, and where do we stand
06:48right now?
06:49I mean, obviously, people think about the EU, and they think of it as being fractious
06:55and always arguing and disagreeing, and they haven't seen anything until they've seen Southeast
07:00Asia, right?
07:02Much more complex, diverse region, and harder to come together.
07:06So where do we stand?
07:07I mean, I don't know how many of you here are technologists or from the tech sector.
07:12You know how fast technology develops, and then you've worked and interacted with governments.
07:17I was from the public sector.
07:22Governments, public sector don't move that quickly, and with AI particularly, the break
07:27net speed at which different companies are racing to get ahead, get towards AGI, governments
07:32are never going to keep up, let alone a multilateral institution like ASEAN.
07:37So you need a framework, but you also very quickly need to develop meaningful public-private
07:43partnerships, and that can help forward the goal.
07:48So that's, I'm going to pause there, because I can talk endlessly about this.
07:51No, that's really critical, and it's interesting, I mean, the Singapore story definitely underscores
07:56that.
07:57But do either of you have kind of thoughts about how far along we are in terms of discussions
08:02about a framework?
08:03Well, so I'm a glass-half-full kind of person, right?
08:08So because I look at it from a perspective that, number one, ASEAN is not like the EU.
08:15The way we are put together is very different.
08:17So we can't expect the cohesiveness of EU in terms of policymaking.
08:23Second, imagine if we didn't have ASEAN, right?
08:27So while there's a lot that can be said for moving faster, for this kind of organisation,
08:34of association, of nations, this is the speed that everyone's comfortable with.
08:40It's better than not having something.
08:42But I would say, for AI governance, it's likely to take the same kind of pathway that ASEAN
08:52took for data protection.
08:53You just imagine, back in 2016, the ASEAN ministers for telecommunications adopted a
09:01set of data protection principles for ASEAN.
09:05Progressively, over the years, they built out the data management framework, model contractual
09:10clauses to assist in the cross-border transfer of personal data.
09:17It takes time, but with concerted effort, things can be developed.
09:23Now these are still voluntary frameworks, they are not binding on the governments, but
09:27at least they exist.
09:29And once they exist, and it's put out by an organisation like ASEAN, there is a bit of
09:34impetus to say, let's make use of them and let's try to be coherent.
09:39I can imagine that for AI, the same path might be taken as well.
09:44I think the first step has already been taken, the responsible AI framework has been set out.
09:51So I think we can expect more of this down the road.
09:55A lot of work still needs to be done by us in the room, as business owners, to really
10:02reinforce that message as we are talking to policymakers and governments, to say that,
10:08look, ASEAN's put out something, can we just make sure that when you introduce national
10:12policies, be as coherent as possible, use it as a mapping standard, because it helps
10:19businesses make sure that their internal compliance programme is easy to manage.
10:24It makes it a lot easier to deploy products and services across Southeast Asia.
10:29So I don't claim to be an expert on the ASEAN discussions about this, but I was struck to
10:36notice that there was a big delegation of folks from the EU who trooped to Southeast
10:41Asia last year and tried to persuade everyone to, let's link up, let's have a common standard
10:47for both our markets, not just one market, and to adopt our GDPR framework and then it'll
10:55be easier for companies.
10:56And it makes a lot of sense, but the response was not very enthusiastic from Southeast Asia.
11:03And surprisingly, what the Southeast Asians said is actually we're very suspicious of
11:08over-regulation because we're afraid that we're going to miss out on some of the benefits
11:12of innovation and these technologies will pass us by.
11:16So should we be discouraged by that, or in a way, can we be encouraged by it if we think
11:22that at least in that regard, all these countries seem to have sort of a common outlook?
11:27There seemed to be something resembling a consensus on that.
11:30Do you have any thoughts about that?
11:33Loads, I'll jump in, but I would love to hear from my colleague here too.
11:39We're sort of a mix, a hybrid region, right?
11:44We're very diverse.
11:46We love each other, yet at the same time, we're always looking over our shoulder and
11:49wondering what our neighbours are doing.
11:52So our model has always tended to be a hybrid of both, regulate when we need to regulate
11:58or we want to regulate, and regulation tend to come fast and quick here.
12:02So we always need to keep a sense of the pulse.
12:04But at the same time, we like to keep an ecosystem that's open to innovation to test things out.
12:14And we like that, because what you see happening in the EU is regulation comes up pretty quickly
12:20and to a certain extent, it stifles innovation.
12:22So how do you do this balance?
12:24And I like what's happening in Thailand.
12:25I like what's happening in Singapore, where governments are reaching out and thinking
12:29about how do I partner up with the private sector to think about meaningful applications of AI?
12:35And I agree with my panellists here.
12:38In Southeast Asia, we're at the stage of application, not necessarily building up the technology
12:43from ground up per se.
12:44So governments really need to think about what makes sense for my country and to ensure
12:51that the infrastructure available in the country supports the application.
12:55So for example, in Thailand, Thailand 4.0, for example, looks at industry application of AI.
13:01And you know how they're gunning for a lot of EV investments in Thailand and making use
13:05of AI to automate certain technology, certain part of their manufacturing is putting them
13:11ahead of the game.
13:12Yeah.
13:13I mean, Thailand is kind of streets ahead of almost everyone in the region in terms
13:17of adopting solar, I think.
13:19Yeah.
13:20And also in terms of data centres, for example, like Malaysia is gunning for tier three data
13:25centres to make sure that cybersecurity integrity, electricity availability supports the ecosystem.
13:32And similarly, in Singapore, we lifted our moratorium in late 2023, but placed very high
13:37standards for the next generation of data centres that's opening up.
13:41So really, government's role here is to partner up with private sector and to ensure that
13:46ecosystem supports innovation and development of enterprise.
13:49So I want to dig a little more deeply into that, because I think there's a lot to unpack
13:54there.
13:55But Evie, do you have a thought about it?
13:57I think just building on what you were talking about, Clay, with the campaign from the EU
14:02to sort of do something similar around AI here in the region, as they did prior to GDPR.
14:08I think that where enterprises are trying to do something different is looking beyond
14:13concepts like being a good steward of the data and thinking about how AI systems interact
14:18with each other and the unique risks related to AI.
14:22And for that reason, there's a little bit different approach taken to the regulatory
14:25aspect.
14:26So the EU has the ability to be united in its enforcement of the EU AI Act, which I
14:32think is a big differentiating factor for why they're approaching the regulatory with
14:35AI.
14:36Right.
14:37Yeah.
14:38You've got to have a much more kind of consensus approach here.
14:39Yeah.
14:40So there was quite a little agenda for that trade mission you mentioned, right?
14:44So while they came and, as you mentioned, they wanted to evangelize GDPR, there wasn't
14:53any take up, right?
14:55But what came out of that?
14:57What came out of that was actually a joint effort between the European Commission and
15:04ASEAN.
15:05And they came together and they did a mapping of the standard contractual clauses and model
15:11contractual clauses that came out of ASEAN, standard contractual clauses that came out
15:15of EU, because they recognized that trade needs to carry on between these two regions.
15:21So let's have some effort to try and map and make it a lot easier for businesses.
15:27So while there's a recognition that maybe some of these policies are not entirely suitable,
15:33there is a recognition that trade needs to flow, data needs to flow, and it's that kind
15:37of spirit, I think, that we need to encourage.
15:39Interesting.
15:40I want to leave some time to take some questions from the floor.
15:42If anybody's got burning queries, put your hand up and wave at us and we'll recognize
15:48you.
15:49You mentioned the sort of things that many of these Southeast Asian nations have in common
15:55in terms of what they're looking for as a regulatory framework.
15:58Of course, there are individual specific issues that relate to AI that often kind of put them
16:04at odds against each other.
16:06Data centers is one.
16:07You know, as soon as the moratorium happened in Singapore, all the operators rushed across
16:13to Johor and started, you know, in Indonesia, started putting up data centers there until
16:17there were more data centers across the border than there were here.
16:19And then the government went, hmm, maybe this is not what we want.
16:24So that's kind of a Frenui's example.
16:28The other one that's fascinating to me, and I don't know, maybe none of you want to comment
16:31on this, but I spent a little bit of time in the Philippines, and the Philippines is
16:36very ambivalent about AI because of the job implications, right?
16:42You've got a very big population of people who are employed working at these call centers,
16:49and that's probably going to be one of the first places where companies are going to
16:53figure out that they can roll out, you know, interactive chatbots with generative AI.
16:59And so they're much less enthusiastic about it.
17:03How do you see that?
17:04Are there similar tensions in other economies as well, and are those insurmountable?
17:09There's a really good research piece done by Tricia at LinkedIn where they kind of did
17:15a study on the extent to which AI was going to disrupt the employment of an individual
17:22or displace or augment, or it's completely non-implicable, right?
17:28And this study was done on Indonesia.
17:30So for the bulk of the population, AI would augment, or has the potential to augment.
17:37So what governments, since this panel is about government and governance, what governments
17:41need to start thinking about is how do I support my workforce in thinking about how to effectively
17:47leverage AI, and similarly for enterprises.
17:51We actually did research on AI in the BPO sector in the Philippines, and yes, there's
17:56a potential for massive displacement of the workforce.
18:01And so it's about how quickly can they go further up in the packing order of sales sets, right?
18:08So it's a challenge.
18:10What you do see happening is Southeast Asia is still in that stage where we're figuring
18:13out adoption of AI.
18:15So we've got a bit of breathing space here for governments to think about how to effectively
18:19manage this.
18:21I think even beyond the comments you've made, thinking about how AI can supplement the existing
18:26workforce or better enable people to do even more beyond what they are currently doing,
18:32and the integral part of that is designing AI in a responsible and trustworthy way.
18:36If an enterprise is able to accurately see who is using AI at the company and how they're
18:41using it, they can actually maximize the benefits of things like generative AI, understanding
18:46and sharing prompts amongst employees so that you know, this is how my enterprise is
18:50using it, this is what's working for my employees, this is what's not working, and putting your
18:55own guardrails around it, I think is a really important part to it, augmenting that workforce
18:58as opposed to replacing it.
18:59So a lot of this is it creates communication challenges and responsibilities, not just
19:04for companies, employers, but also for governments to say, oh, look, this is going to be net positive
19:09for all of us.
19:10There are going to be some winners and losers, but trust us, we're going to do what we can
19:14to try to adjust and adjudicate that, give people the skills they need.
19:18That's hard, right?
19:19That's hard to do.
19:20Well, I'm going to just add to say that while I can't speak for Philippines and call center,
19:26I am quite involved in Singapore and legal sector, and legal sector is one of those that's
19:32expected to be disrupted by especially generative AI, but I think it all falls on us wherever
19:39we are to actually tick the bull by the horns, right?
19:42So the Academy of Law has articulated a four-pronged approach to this.
19:48Number one, we actually need to just grapple with this and find ways to teach our lawyers
19:53not only to use AI, but also to teach them differently because AI tools are going to
19:58make their way into tools that will help them become more productive and tools that their
20:03clients are going to use.
20:05We need to work with solution providers to get AI solutions into the market so that they're
20:10using it, right?
20:13Of course, for us, we also want to make sure that the lawyers learn how to do things like
20:20problem engineering, and of course, for our own products and services, we need to also
20:25use AI to enhance our feature set.
20:28So it's really up to us, wherever we are seated, to try and get something done and not just
20:35wait for the disruption to happen, but let's try and get ahead of that.
20:39Sounds like some thorny challenges.
20:40That is, unfortunately, all the time that we have, but please join me in thanking our
20:44panelists for a really fascinating discussion.

Recommended