Josephine Teo on AI at the Bloomberg New Economy Forum

The session opens on the global AI divide. UN data shows over 100 countries lagging on adoption, while 30% of US internet users still cannot stream Zoom reliably. Google DeepMind COO Lila Ibrahim says models must be designed with the world in mind from the start — DeepMind partners with AI Singapore to fine-tune for Southeast Asian languages, and Gemini's multimodal interface lowers the barrier to use. Minister Josephine Teo argues that few countries can own the full AI stack, but every country has agency on capability — workforce skills, enterprise capability, academic strength, and AI governance and use inside government itself.

SMEs are 99% of Singapore's GDP and 70% of its workforce, and they say GPUs and cloud are too expensive. Teo gives an example: a small eatery used an AI assistant inside an existing platform to analyze past sales and plan promotions — no GPUs needed. Manufacturing is roughly 20% of GDP, and a new AI Centre of Excellence for Manufacturing has built a shared model for predictive maintenance on CNC machines. Ibrahim notes Gemma and AlphaFold are used by over 3 million researchers across 190 countries on food security, pollution, and disease. But financial institutions and IP-sensitive organizations cannot rely on open models alone — the future is a mix.

On education, Ibrahim says the real value is personalized learning, not just freeing up teachers. She talks about her dyslexic twin daughter using AI to express ideas she previously couldn't structure. A six-month pilot in Northern Ireland with 100 teachers found average savings of 10 hours per teacher per week. Both speakers flag harder questions: what does it mean for community and family if a child's best friend is an AI chatbot? DeepMind keeps ethics researchers and child psychologists on staff to address these questions before they harden. In the audience poll, the top driver of AI-driven inequality was "concentration of power in a handful of companies and countries"; mass job displacement ranked last.

Caption language: en · Fetched: 2026-05-02

tech revolution of our time and with it comes trillions of dollars of economic value or so they say. The thing is according to the United Nations more than a hundred countries around the world are lagging way behind. They're not even close to adoption and many of them are in the global south. The other thing as well is that within the US, 30% of those using the internet are using it at speeds not even fast enough for Zoom. So essentially, billions of people around the world are at risk of being left behind in this AI revolution. So what gives? How do you bridge that gap? And how do you bridge that gap fast enough? because this is a a technology that's moving really fast. Laya, let's start with you. How do you get AI access to people? As many people as possible, do it at scale. Starting with the easy questions.

[laughter] Um, we have 20 minutes. Uh, you know, this is actually something that is really important. We have to make sure as we're building AI that we're thinking about the global implications of this technology. Part of this means how do you uh develop the models as well with the world in mind. It's not something that we can think about as an afterthought. Um for example, one of the activities we've been doing here um in Singapore is working with AI Singapore to ensure that Southeast Asian languages are well represented. And that means taking one of our open source models and fine-tuning it uh with you know getting the data right first then fine-tuning the model and making it available uh to a broader audience. To me this is you know AI is not a technology where we have national borders. You launch a model and it's available worldwide.

So we have to think about everything from access to literacy to actually making the models easy to interact with things like making it multimodal from the beginning um which we've done with Gemini which is you can talk to it you can type it you can take a picture and being able to do that in many different languages being able to transform that output so you're lowering the barrier for people in general to use it. There's still a lot of work to do, but we're, you know, we're we're pleased to be making uh this progress and it's only happening because we're doing it in collaboration with local ecosystems that are helping with that like last mile access. But massive barriers, right, Minister infrastructure, AI, infrastructure, it's key, compute capacity, those things will decide and shape who really gets full access to AI.

Well, I think if you characterize it as AI halves and have nots, um, very few countries in the world will be able to have all the elements of a AI ecosystem that are desirable, that are good. It's a given. It's a reality. I don't think it's going to change. But I also think we are not without agency. uh within each of our own countries, we do have the ability to reshape things and some of the areas that are worthwhile reshaping in order to be more ready for an AI future is investing in infrastructure making it happen. You may not be able to do it on your own, but there are organizations like the World Bank, IMF, even the United Nations Development Program that are very alive to this problem and leaning forward to help.

Um, at the same time, you also want to think about creating the right conditions for the technology to be used in a responsible way. So at the right time putting in place guardrails is something that every country can think about doing. But to my mind the single most important area of intervention of worthwhile effort and where you must exercise the most agency is really in developing capabilities and the capabilities have to be thought of in a holistic way. capabilities in the workforce, capabilities in enterprises, capabilities within academia, actually even capabilities in AI governance and AI use within government itself and different kinds of capabilities, entrepreneurial capabilities, engineering capabilities.

There is, you know, a whole range of things that need to be built up in order for a country to enable its citizens to benefit from AI and to have the best chance of moving ahead. Will you always get to what the next country is able to achieve? Maybe yes, maybe no, but that shouldn't hold us back from trying. >> We can try and access. If you take a look at what we're hearing from the small and medium enterprises, 99% of the GDP in Singapore economy, 70% of the workforce, they're saying that it's just too expensive. The rising cost of GPUs, cloud capacity, we know Google has taken that into consideration. How are you engaging withmemes to counter the difficulties that they're going through? Well um you know within Google DeepMind work the AI uh innovation engine for Google products.

So one of the things that we really uh are working on with our Google our colleagues across Google is how do we integrate um our AI tools into existing programs. You know in some ways I wish we could actually rebrand AI like it should if it's a good feature and it's actually being useful it's just part of an everyday workflow. So the first step I think is like actually getting it integrated and removing that barrier. But there is something about um reducing the barrier of like the literacy aspect. And part of this is it's been quite amazing to see um the past few years how many different resources are available uh in order to be able to use the technology. So one is like just getting it integrated.

The second is removing the barriers as we talked about in terms of just um you know making the context window which is the prompt window uh larger so that you can input a lot of different information in a natural way and then the other one is making sure that the users actually have the skill sets to do it. Minister >> well I thinkmemes are very heterogeneous bunch and there is no one single way of characterizing them and in terms of their adoption likewise there will be different pathways I'll tell you a story it was quite um you know encouraging for me uh this is a little hole in the wall along road they are a small ery and um I happened to chance upon them one day and I asked asked him how business was and I we ended up talking about AI and they told me something quite unexpected. They said AI has been helping them do better business.

I said how? He said well you know I have to run promotions and I've been running promotions in the past years. Some are more successful, some are less successful. And um as it turns out, I now go to the platform which has an AI assistant I can use and I can ask it based on my past sales record to tell me which of my past promotions have been more successful and that has helped me tremendously to plan the next promotion. Uh the point of raising this example is to say that not every single theme is going to have to to need its own access to GPUs. In this case, it's not even GPUs. It's just an app that was made available to this little eery and an AI tool that was made available through it. there will be othermemes where being able to use um a fine-tuned model uh could be meaningful.

If you take the fact that uh Singapore's GDP still has about 20% contributed by manufacturing and many of thesememes perhaps engaged in precision engineering work. We now have an AI center of excellence for manufacturing that has recently developed a common model that does predictive maintenance for CNC machines which many of thesememes will use. So the point I'm saying is that uh really there will be different methods of us being able to supportmemes. Some will come through tools that are made available by bigger platforms. Some will be through efforts that are targeted at the sector, targeted at the kinds of companies in them. I think it's more important that we look at how we can make meaningful differences to how the the way businesses are operated.

It is still about leveling the playing field and there are different options out there. If you take a look at China, it's open source. You know they say innovation is faster is cheaper and then you have closed. How do you weigh the pros and cons of the two? We understand the concerns of a data security cyber breaches but there are pros to an open system. Yeah. It's uh it's something that we have a long history in actually making the technology available from the early days of our the research into um open source models.

uh whether it's something we call Gemma or even Alphafold for the scientific community to make it broadly available with um we have 190 countries over three million researchers using this to advance a number of fields whether it's uh in food security pollution un uh dealing with industrial waste or even understanding diseases to advance therapeutics and in many ways that in of itself is lowering barriers because now it's democratizing access to this information but we also have to think about uh every case is different. So this balance of how can we uh seek the benefits while also managing the risk and as the large language models and the mult uh multimodal models get more capable this introduces a new level of risk.

So when we released alphafold our uh protein prediction model we actually sought the outside expertise of scientists academics Nobel laureates in this space to say what are the risks we've assessed it this way what do you think what's the right way to deploy the technology so it's kind of on a case-byase basis as AI gets more general and capable this is something though that we're doing extra testing so we've just released today Gemini 3 which is are not only our state-of-the-art most advanced model but also the most tested and I think that's part of the pace of the technology right now because we still have issues with factuality reasoning understanding the context these are things that are not going to be solved right away.

So it's this balance of how do you make sure the technology is available so people can build off of it they can advance their businesses their uh economic opportunity learning etc while still managing the risks >> I think if you're talking about widespread adoption um having access to open models are very important and um you think about the thousands and perhaps in in time to come millions of applications that uh um very wide variety of organizations are going to have to use and individuals are going to have to use without access to open models. I think that would be quite problematic. On the other hand, um if you're talking about a financial institution, if you're talking about an organization that has a lot of intellectual property or security concerns, expecting them to only rely on open models is clearly not going to be feasible.

So I think the future will see a mix of these different kinds of uh access being made available and we must do our part to try and ensure that access remains open as widely as possible if we want to see widespread adoption. >> I'd like to touch on education but before that there is a poll would love for everybody to participate. The question is how will AI most exacerbate global inequality? A mass job displacement, B concentration of power in a handful of companies and countries. C insatiable demand for energy and D widening capability gap between advanced and developing economies. While you decide on your choice, let's talk about education. You know minister we talk about how you must prepare the next generation.

How are you thinking about the possibility of the elite institutions so wealthy countries uh being the next I guess source of AI talent? Is is there risk of that happening? Well, there's definitely going to be competition to try and get as much of an advantage as possible in terms of raising the quality of human capital in each of our countries. And I don't think you can stop anyone from doing that. Um this sort of effort um will take place with or without um anyone you know putting a stop to it or trying to boost it. I think what people um can attempt to do is to ensure that they themselves have the ability to learn. And I think that in our education system, we do want to make sure that people uh children have exposure at a young age but in an appropriate way.

Um the way AI is being used in our education system I think is being experimented upon. At one level it was about enabling the teachers to be more productive to be able to outsource as it were the more mundane aspects of their work perhaps administrative work to an AI assistant. Uh that is not in and of itself harmful but it also kind of misses the point um of uh you know using AI for education. The great value of AI is in personalizing learning in a way that attends to the unique strengths and shortcomings of every learner and trying to find ways to enable them to achieve better learning outcomes.

So um I'm very glad to say that uh my colleague the minister for education and I have had deep conversations about this and our ministry of education is looking at it not just as a way to support teachers but more as a way to support learning and that is the direction I think we all should try and move towards. Now is it something that every education system is able to do? Um I I think we were just talking about this morning the fact that Google Deep Mind has set up a research lab in Singapore and one of the potential areas for collaboration is identifying ways in which um learning education can be made more effective and since we said in Singapore AI for the public good for Singapore and the world um that being our vision we hope that the benefits don't just acrue to students in Singapore but potentially to students elsewhere.

And to that point, how do you cater to the underserved? I mean, you're an advocate for education and reskilling and skilling among such communities. What's your take on that? >> Well, u yeah, education is something that's really important to me. And just to to comment on a few things there, um I think that you have in many ways starting with teachers is good. the productivity. Uh we're seeing around the world that teachers are really struggling to keep up with the workload as class sizes may get larger, as different skills may emerge within students. We just did a pilot for example in Northern Ireland where uh with the government there um and 100 teachers for six months and they found on average they reported 10 hours of savings per week per teacher. That's pretty significant. And what does that mean?

And it's not just about saving time, but it's actually letting the teacher do what they wanted to do in the first place, which is transfer their knowledge and passion for a topic to the students, which ultimately drives learning outcomes. So if I think if we can come at this uh AI and learning from a different perspective, which is yes, you have to start with the teachers, but the teachers are the magic in the classroom. So how do you actually free them up and then give them the tools to help with the learning outcomes and also meet the learners with where they're at in their learning? We all learn differently. Uh whether you're a visual learner, you prefer to listen to podcasts to learn.

So I think now that we have AI where it's uh again multimodal from the ground up where you can now interact with a content in a different way, you can learn about physics by learning about your favorite sport and in the context or instead of just giving an answer, being able to go multi-turn with the the model. This is the type of technology that we think is really important. It all starts with treating learning like a first class science problem, right? It's not just about physics or biology. It's like uh which AI uh at Google Deepmind we tend to focus on. It was actually investing in pedagogy uh and knowing that we come at it from an AI expertise. So who are the learning experts that we can go and talk with and make sure that our models and our AI development is grounded in this? We need to make the AI more factual.

We need to uh make sure that it has the best of learning science uh built into it, but we also need to work with the community because at the end of the day, this isn't about just layering AI on top of the existing education infrastructure. It's also rethinking about how are we going to train the next generation and how are we going to ensure that they have this as a tool that they use responsibly. They're not just offloading uh the learning but really engaging with it. So I think this is an area where one is we're still very early, two is it's extraordinarily high priority and then we think uh we really believe that here in Singapore especially the um extraordinary history of um workforce training of lifelong learning of education as an essential value of the culture is really critical.

It's an area we're really looking forward to collaborating more >> in terms of exposure. I was really surprised. I was in Vietnam not too long ago and there was a school exposing kids as young as six years old to AI. How do you think about, you know, when is a good time, how early should it be for for students and kids to be exposed to AI? I mean, what's the case in Singapore? What's the thinking? >> Define exposure.

uh just being able to >> looking at AI generated content is also exposure [laughter] >> or >> more than that and >> perhaps as well right >> so I mean how do we look at when you talk about education how early should it be and I'll be curious to hear as well from Laya's perspective in terms of what's key what's the one attribute what's what what do we need for an AI centric future >> yeah well maybe I'll answer this as a parent rather than just a an a leader in the AI space which is I have twin daughters and my twin daughters are my natural AB experiment every single day I have teenagers which [laughter] uh yes thank you that's how I feel sometimes >> it turns out I have twin daughters >> oh you do ah and we both wore purple um [laughter] uh so actually yeah my twin daughters opportunity to see and and and I think there is this question of age like what's the right age but around the age of eight we learned learned that one of my daughters was dyslexic and the other daughter is a traditional learner.

And I saw my daughter go through the the uh English school system because we're based in I'm based in London and and lose her confidence and start to label herself of like not being good at school, not being good at certain subjects. And it was horrible as a parent. And but what I've seen as she's gotten into her teenage years is that she can now use AI to help express herself in a way that other people will understand her. You know, the ideas were always there, but she just didn't know how to structure her thinking in a way or maybe she can now uh take some type of text and listen to it as a podcast so that she can learn. So I think it's less about age and more about putting the right scaffolding in place to really bring out the human potential.

Now whether that's a student, whether it's a worker, a small business owner, me as a leader in my time poor life, but I think it's this kind of gets back to like how are we not just about the access but being able to provide the right infrastructure and and in in maximizing what we can have to offer into the world. I would say that the question about um the impact of AI on education opens up um so many um unanswered you know themes. For example um a very central objective of education is to enable people to earn a living by performing tasks that they are trained for. We do that in you know in in in schools and even after a person leaves schools we have continuing education and training and that's really for the purposes of a per person being able to hold a job that pays an income.

Now if more and more of those types of work can be done by an AI and less and less require the kind of training that we have put putting in to prepare individuals for then we have to come back to the original question of what is the purpose of education what do we train or what do we help people to learn so it's not just learning outcomes it's what do you learn in the first place >> and I don't know that there are easy answers to this question. But I do think that this is an area that is very well worth our effort in investing in research in understanding more deeply uh about the impact uh of AI on human societies and how we need to interact in human society. So for example, if a child grows up learning that the best friend is an AI chatbot, what does that mean for communities? What does it mean for family bonding?

And what does that require us to do in either to lean against it or to help the individual understand the use of this AI in the context of his daily interactions? I don't know that there are easy answers, but I think that these are questions that we have to work together on. And by the way, if I could just say, I think there are questions we have to ask and have open conversation and debate, collaborate on and then we have to figure but we have to ask those questions before it's too late, right? So we have a team of ethics researchers of ch a child psychologists so that we can actually how do we think about building this in and it's a it's not a oneshot deal. We have to continuously improve. Before we wrap up, I just like to get back to the poll. If we can get the uh I know it was flashing for a very long time, but I didn't get to it.

If we can get the results of the poll again, how will AI most exacerbate uh global inequality and most of the um participants say it's concentration of power in a handful of companies and countries. >> What stood out for me was actually that mass job displacement was ranked very lowly and these mainly being employers. Is that something to [laughter] take away? >> Actually feeling more hopeful about >> are you surprised >> employers are not thinking that there will be mass job displacements is that a good thing that must be a good thing. But if you say that they're full of employers here then perhaps it's um it's only to be expected.

But the concentration of power in a handful of companies and countries that is concerning right I mean if you were to refrain I mean is there a way of refraraming this discussion so that people can I guess counter the concern of the concentration of power? How would that be? I mean how do we address that particular issue? Um from my perspective on that one is I think uh I'm as mentioned by the minister I'm really excited that we are opening an our first research lab that Google DeepMind has ever opened from the ground up here in Singapore. So I think that's going to be really important for thinking about a more globally distributed uh presence and making sure that the representation of culture and languages research areas are are integrated.

I also think it's really important that companies are collaborating uh and really thinking about the ecosystem about uh engaging in a way that enables rises all tides and so that having this as an open debate actually holds us all to account which I think is really critical. Yeah, I would just add that um it is a serious concern because there are so many things in AI that requires collaboration. For example, the setting of standards to make sure that this AI doesn't end up, you know, harming human interests in a very significant way. And each country uh feels that it has a stake in this conversation and does not want to be left out of um you know the rules making that uh will eventually have to come. we hope and and and that's why we participate actively in international fora and um we also have conversations with our colleagues.

Singapore as you know is the convenor of the forum of small states and you know some years ago we decided to set up a digital pillar. There's been tremendous interest in how we can advance the interests of countries that are smaller in size, but who also want to benefit from this technology and who also want to make sure that their citizens don't get the short end of the stick and it's still an ongoing conversation. Um, it's hard to say whether we should be hopeful or not, but it's an area that we have to pay serious attention to. >> It's still a work in progress. Minister Tio Laya, thank you so much for your time today. Thank you. >> Thank you. >> Thank you. Thanks. >> [music] >> Heat. Heat.