Keynote Address by Minister Josephine Teo at the Asia Economic Summit in Jakarta, Indonesia

1. Selamat pagi. I’m happy to be back in Jakarta, and to see my good friend, Minister Meutya. She was very kind to refer to me in her speech, but I would like to remind everyone that although there are many digital ministers who are women, there is only one that has to look after 280 million people.

2. About a month ago, Minister Meutya and five other ASEAN digital ministers were in Singapore for the Asia Tech Summit. This is one of the many occasions we have met. It speaks to our shared interest in advancing digital development, in each of our countries and in ASEAN. We are always asking, “what more can we do together?”

3. I will focus on two things today: First, why Southeast Asia is well-positioned to expand AI adoption, and; second, the pitfalls we must avoid to do so confidently.

 The Case for AI adoption in Southeast Asia

4. First, let me start with the positives. In every region, achieving widespread AI adoption will likely depend on five key building blocks.

   1. The first, reliable infrastructure;

   2. The second, capabilities within organisations and the workforce;

   3. Third, whole-of-society involvement;

   4. Fourth, supportive laws and regulations; and

   5. Fifth, international cooperation.

5. In ASEAN, we’re firing on all five cylinders.

Pillar 1 – Infrastructure

6. The infrastructure build-up is considerable

   1. New data centre capacity is coming online almost everywhere,

   2. Between 2025 and 2030, capacity is expected to more than triple. If I share with you this morning’s conversation before the start of this summit, triple is more than modest. The number sounded like more than ten times.

   3. Our seabed is fast becoming one of the world's busiest networks of digital highways.

7. However, as Minister Meutya correctly pointed out, infrastructure is more than just cables and computing power. It includes the software and policy layers that make AI accessible and useful.

   1. For example, how well do AI models and applications respond to our cultural diversity?

   2. SEA-LION, which stands for Southeast Asian Languages In One Network, is one software layer that can help. As an open-source model, anyone can use it freely.

   3. Today, it has over 200,000 downloads. Many companies across the region build their products on top of it.

Pillar 2 – Capabilities

8. Beyond infrastructure, organisations and people must know how to use AI. In fact, more companies are integrating AI into their operations. That is actually the more difficult thing to do, and Minister Meutya was talking about it earlier.

   1. Banks are using it for fraud detection. With AI, manufacturers can better optimise supply chains and reduce downtime.

   2. Companies like Kata.AI are creating AI tools to enhance customer service in Bahasa Indonesia.

   3. It is timely that this year’s ASEAN Chair, the Philippines has prioritised helping Micro, Small and Medium Enterprises (MSMEs) to adopt AI.

   4. In parallel, the ASEAN Foundation plans to empower 100,000 MSMEs to use digital tools.

9. Our people are young, digitally connected, and quick to adapt. Many governments are looking to equip their students with AI-ready skills and partnering with tech companies to level up their workforce.

Pillar 3 – Whole-of-Society

10. This brings me to the democratising value of AI access. We saw this clearly at the recent AI-Ready ASEAN Youth Challenge.

    1. Two university students from Brunei built Sahabat-Care – an AI platform that supports dementia patients and their caregivers.

    2. Indonesia’s Noah AI tackled a different problem. They built a platform to predict, verify, and send SMS alerts about floods in real time.

11. Like many innovators today, these young creators worked around local constraints, such as low mobile connectivity in the far-flung areas, sometimes 3G or less. These ground-up initiatives spread AI’s practical value to under-served communities.

Pillar 4 – Governance

12. The wider the reach of AI, the more we need good governance to manage the risks and stay accountable.

    1. Some AI risks can be addressed by existing laws; for example, employment laws may already cover workplace discrimination caused by problematic AI tools.

    2. In other cases, such as the use of AI-generated materials during elections, the law may need to be updated.

13. As the science is evolving, it is sometimes more useful to set out the principles, rather than to prescribe the regulation. This is the approach we have adopted in Singapore.

    1. We were the first country in Asia to launch a Model AI Governance Framework.

    2. We have also published guidelines on responsible AI.

14. ASEAN has adopted broadly the same approach through its Guide on AI Governance and Ethics and its Programme on AI Governance and Safety Testing.

Pillar 5 – International Cooperation

15. These were possible because we already have an ASEAN Working Group on AI Governance. Looking beyond ASEAN, we have also been working with technology partners to build capacity.

16. We are active at international platforms like the AI Impact Summits or the ISO, ensuring that our region has a seat at the table, and a voice to shape the standards around AI.

Potential pitfalls

17. Overall, Southeast Asia is in a good position to deploy AI widely. But we will not get there automatically. Every now and then, it is useful to talk about what could derail our best plans. I want to name two today: data and sovereignty.

Pitfall 1 - Data

18. The first potential pitfall is our data policies.

19. Data is the lifeblood of AI.

    1. To train reliable and robust AI models, we need access to quality data, and a supportive environment for data to flow.

    2. For example, businesses may need a holistic view of their interactions with customers and suppliers in multiple countries, to plan and optimise their operations. If they operate only in a single country, that is not a problem, but if they expand and scale beyond one territory, then the interoperability of the data policies becomes very possible.

20. Some restrictions are of course necessary to protect personal and sensitive data, or national security. But for the wider economy, onerous restrictions on data hamper innovation and business, especially for small companies that do not always have the resources to overcome compliance hurdles.

21. Collectively, we should help these small companies to thrive and scale, whether in Jakarta, Bandung, Hanoi, or Bangkok. This is precisely what the Digital Economy Framework Agreement, or DEFA, aims to achieve.

22. DEFA is a landmark agreement that will put in place common rules and frameworks to enable digital trade, and support trusted cross-border data flows.

23. Several countries were key in shaping DEFA outcomes. I want to acknowledge colleagues in Thailand, the Philippines, and Indonesia. They have laid a strong foundation for ASEAN’s digital economy to grow.

Pitfall 2 – A Narrow Reading of "Sovereignty"

24. The second potential pitfall is how we respond to the need  for AI sovereignty.

25. Sovereignty is a legitimate concern and rightly so. Every country must be able to use AI on its own terms, in service of its own people. At times, this question gets framed narrowly as one of ownership or a race to build or claim full possession of the entire AI stack: chips, models, data, and applications.

26. This view, while understandable, is neither realistic nor helpful for most countries.

    1. Setting aside the enormous costs involved, the AI tech stack relies on intricate global supply chains that very few can hope to build or own at every level.

    2. Many of us have other fiscal priorities, such as to take care of our aging populations, healthcare, or education.

27. I suggest that we can frame our goals in three ways:

    1. First, the ability to use and govern AI for the public good.

    2. Second, the autonomy to make smart choices about who to buy from, who to partner with and on what terms.

    3. Third, the anchors to build breadth and depth in our AI ecosystems, such as our research institutions, developer communities, venture capitalists, and ambitious users.

28. These three goals will already take considerable effort to achieve. But they are practical ways that help us to protect our citizens’ and strategic interests.  They should be our priorities.

Close

29. Next year, Singapore will assume the ASEAN Chair. We will build on the Philippines’ good work to advance our shared AI priorities.

30. Here is what to look forward to:

    1. First, we will bring more MSMEs, workers, and governments together to use AI better.

    2. Second, we will invest more in shared digital public goods, such as language models, governance toolkits, and capacity-building programmes.

    3. Third, we will deepen cross-border data flow mechanisms and align AI governance approaches across our region.

31. This is not just Singapore’s agenda. This is ASEAN's agenda.

    1. ASEAN’s strength has never been about being the same.

    2. It has always been about working together, despite being different.

    3. Indonesia embodies this. As your own national motto puts it: “Bhinneka Tunggal Ika” or unity in diversity.   

    4. As long as ASEAN moves ahead in step with one another, we can achieve much more.

32. I look forward to the panel discussion, and to learning from all of you. Terima kasih.