MDDI 演讲稿 · 2026-05-20
杨莉明部长在 ATxSummit 2026 上的开幕主题演讲
杨莉明部长在 ATxSummit 2026 上的开幕主题演讲
要点
- • 新加坡建立了七十个AI卓越中心。
- • 国家AI影响项目旨在帮助一万家中小企业有意义地使用AI。
- • 连通性、先进制造、医疗健康和金融四个部门被列为国家AI任务的优先领域，合计占新加坡国内生产总值的40%以上。
- • 樟宜机场第5航站楼将使该机场的年度客运处理能力增加约5000万人次。
- • 新加坡的工业机器人密度是全球平均水平的约五倍，始终排名全球最高水平。
- • 由总理主持的国家AI委员会已成立，旨在监督国家AI任务以扩大AI在经济中的采纳。
完整译文（中文）
MDDI 英文原文译文 · 翻译日期: 2026-05-21
尊敬的诸位，
尊敬的同事和朋友们，
各位早上好，感谢各位的莅临。
三年前，在这个舞台上，我谈到了我们「人工智能造福公众、造福新加坡、造福世界」的愿景。
那年晚些时候，总理劳伦斯·黄宣布启动《国家人工智能战略》(NAIS) 2.0，该战略概述了充满活力的人工智能生态系统的10个关键促成要素。
快进到2026年5月，全球人工智能发展步伐加快。
在新加坡，已经建立了70个人工智能卓越中心；
我们开发了反映区域特点的开源模型SEA-LION和MERaLiON；
我们的「科学人工智能」项目正在应对疫苗开发等紧迫挑战。
今年早些时候，我们成立了由总理本人担任主席的国家人工智能委员会。
该委员会将专注于国家人工智能任务，以扩大在我们经济中的采用范围，建立深厚的人工智能能力，并使新加坡成为人工智能创新的领先枢纽。
为了更好地支持委员会的工作，我们正在更新我们的战略和优先事项。
这个更新基于我们在实施NAIS 2.0过程中的经验和洞察。
这是一次「深化」而非系统重启。
我们正在细化如何在人工智能生态系统的各个组成部分中突破界限。
以产业为例，小规模项目的生产力影响只是开始。
国家人工智能影响计划旨在扩大企业用户的基础。具体来说，我们将帮助10,000家中小企业有意义地使用人工智能。
人工智能冠军计划将为准备进行更大规模企业级影响的用户提供有针对性的支持。
通过我们的国家人工智能任务设定了一个更高的雄心水平，这些任务旨在改变整个行业。
四个部门被列为优先——信息通信、先进制造、医疗保健和金融。
新加坡在这些部门中都享有全球地位。总体而言，它们占我们GDP的40%以上。
它们也是政府赋能工具（如数据获取和监管沙盒）能够催化人工智能突破的领域。
国家人工智能任务将由值得解决的问题陈述推动，不仅对新加坡，对世界也是如此。
它们旨在成为其各自部门中深入、革命性人工智能采用的「滩头阵地」。
工作已经开始。
在航空领域，我们的樟宜航空枢纽与170个其他城市保持连接，是世界上第四繁忙的国际机场。
市民和游客都享受樟宜的无缝且精心设计的体验。
在未来十年内，新的T5航站楼每年将使樟宜机场的客运处理能力增加约5千万名乘客。
除了建设世界一流的基础设施，我们必须重新思考我们的航空枢纽如何运营。
乘客如何从一个登机口移动到另一个登机口？行李如何在多个航站楼之间运输？飞机着陆和起飞如何在我们的跑道上排序？
这些确实是需要硬件和软件创新的真正难题。
仅有一个新航站楼是不够的。要使我们提供的航班数量翻倍，存在多个制约因素。
在其他挑战中，我们需要一个优先考虑安全和容量的下一代空中交通管理系统。
这只是 AI 可以帮助的航空业众多令人兴奋的机遇之一。
我们在大士港的海事枢纽是另一个很好的例子。它是位于世界最繁忙水道之一的全球最大自动化集装箱码头。
存在丰富的复杂运营数据集，可用于开发令人兴奋的新解决方案。
明天，我的同事运输部长 Jeffrey Siow 先生将与各位分享他的更多愿景。
我们的连接挑战清楚地表明了新加坡如何可以成为世界的生活实验室。
同样，在先进制造业中，影响的范围和规模是显著的。
物理和具身 AI 的发展与我们的制造公司高度相关。
他们已在效率的最前沿运营，以在激烈竞争中蓬勃发展。
新加坡的工业机器人密度约为全球平均水平的 5 倍，始终位居世界最高之列。
物理 AI 可以帮助进行流程重设计的模拟。
更好的数字孪生可以改进预测性维护，减少材料浪费和生产停机时间。
但这并不那么简单。在研究实验室开发的想法在真实工厂环境中并不总是表现得很好。
要看到真实影响，我们需要来自硬件、软件和运营领域的合作者。
这就是为什么我们欢迎英伟达在新加坡的新研究实验室，专注于具身 AI 和高效 AI。
该实验室将与我们的大学、行业合作伙伴和政府机构携手合作。
它加强了英伟达在新加坡的现有基础，深化了我们的 AI 人才生态系统，并开启了创新的新来源。
我们也在开发勿洛数字区作为前沿试验床。多年前，像我一样不太年轻的新加坡人会记得勿洛曾是一个宁静的村庄。
但在今天的勿洛，我们将创建一个综合数据平台，设计真实的测试场景，以及使机器人能够在整个地区安全运营的规则。
机器人可以帮助我们的工作人员加强对目前服务不足的地区的服务供应。
越来越多的行业合作伙伴网络正在使用勿洛的生态系统进行测试和实验。
这些沙箱和合作有助于传播接受度和采用。
单独来看，新加坡的国内市场可能不值得这么多关注。
使我们引人注目的是我们所连接的全球网络以及我们对可信技术采用的良好记录。
像英伟达这样的领先公司选择在这里加深他们的存在，因为我们为他们的全球足迹增加了有价值的东西。
我们期待有更多的人加入他们。
你会注意到我今天分享内容中的一个模式。勿洛的试验床。我们对樟宜 T5 的野心。等等。
在所有这些努力中，新加坡实体将不是被动的观察者，等待解决方案被放在眼前。
我们将积极参与，与伙伴一起解决痛点，强化我们的生态系统，为商业和工业创造新价值。
这些伙伴关系是激活我们国家人工智能使命的两个关键推动力之一。
另一个是治理——我们认真对待这一点，因为我们在新加坡使用的东西必须值得信任。
尽管人工智能发展飞速，但治理方法远未确定。
但随着人工智能在影响人们生活的领域——医疗保健、金融、交通、公共服务——越来越深地融入，所涉及的风险在增大。
来自自主代理、网络安全威胁、对信息的信任侵蚀的风险——它们不受国界限制。
人工智能的好处也可能分布不均。获取机会减少和缺乏包容性将对新加坡这样的小国造成不利。
国际公认的规则和标准将很重要。但它们的形成需要时间。
与此同时，我们已经开始传播和塑造期望，监测它们如何塑造现实情况，并在实践中进行调整。
一个例子是《代理人工智能模型治理框架》，它由行业意见开发，并在一月份的世界经济论坛上发布。
虽然只过了几个月，但我们今天正在更新该框架，包括来自普华永道和Workday等公司的真实代理人工智能部署案例研究。
我们还发布了关于OpenClaw的案例研究，强调了用户的最佳实践。
在政府内部，我们对自己的代理使用应用相同的迭代过程，通过与谷歌等伙伴的沙盒进行实验，并学习如何确保其安全。
同时，我们与全球人工智能安全从业者社区紧密互动。
事实上，周一，一个由Yoshua Bengio教授、Stuart Russell、Dawn Song、Zhang Ya-Qin和Max Tegmark组成的杰出小组刚刚更新了《新加坡人工智能安全研究优先事项共识》。
这些努力承认，在我们能够确保安全设计原则之前，需要大量工作和国际合作，即人工智能在重要领域被负责任地部署。
新加坡将其视为时间和资源的必要投资，我们将长期致力于此，以便人工智能能够被信任。
我概述的这些努力是人工智能在我们的经济和社会中主流化的一部分。
因此，新加坡的人工智能中心日益充满活力。
我们的国家人工智能使命为领先的人工智能公司在此扎根提供了强有力的理由，以开发、测试和扩展值得信任且具有全球相关性的人工智能解决方案。
与此同时，在医疗保健和教育等领域，我们正在利用人工智能造福公众，并改善我们公民的福祉。
我们将在合适的平台上更多地谈论这些内容。
总之，我们希望您在ATx的时间能让您更清楚地了解与新加坡携手构建意味着什么。
我邀请您加入我们，共同创造更大的积极影响。
我祝你们在ATx取得丰硕的成果。
英文原文
MDDI 官网原始记录 · 抓取日期: 2026-05-21
Excellencies,
Distinguished colleagues and friends,
A very good morning and thank you all for being here.
Three years ago, on this stage, I spoke about our vision of “AI for the Public Good, for Singapore and the World”.
Later that year, PM Lawrence Wong launched the National AI Strategy (NAIS) 2.0, which outlined 10 key enablers of a vibrant AI ecosystem.
Fast forward to May 2026, AI developments around the world have accelerated.
In Singapore, 70 AI Centres of Excellence have been set up;
We have built SEA-LION and MERaLiON, open-source models that reflect our regional context;
Our “AI for Science” programme is tackling pressing challenges like vaccine development.
Earlier this year, we set up the National AI Council, chaired by PM himself.
The Council will focus on National AI Missions, to broaden adoption across our economy, build deep AI capabilities, and make Singapore a leading hub for AI innovation.
To better support the work of the Council, we are updating our strategies and priorities.
This refresh builds on our experiences and insights implementing NAIS 2.0.
It is a “double-click” rather than a system reboot.
We are fleshing out how to push the boundaries in each component of our AI ecosystem.
In industry for example, the productivity impact of small-scale projects is just the beginning.
The National AI Impact Programme aims to broaden the base of enterprise users. Specifically, we will help 10,000 SMEs use AI meaningfully.
The Champions of AI programme will provide targeted support for those ready to make a bigger enterprise-wide impact.
An even higher level of ambition has been set through our National AI Missions, which aim to transform entire sectors.
Four sectors are being prioritised – Connectivity, Advanced Manufacturing, Healthcare and Finance.
Singapore has global standing in each of these sectors. Collectively, they make up over 40% of our GDP.
They are also areas where government enablers like data access and regulatory sandboxes can catalyse AI breakthroughs.
The National AI Missions will be driven by problem statements worth solving, not just for Singapore but for the world.
They aim to be “beachheads” for deep, game-changing AI adoption in their sectors.
Work has already begun.
In aviation, our air hub at Changi keeps us connected to 170 other cities and is the 4th busiest international airport in the world.
Citizens and visitors alike enjoy the seamless and thoughtfully designed Changi experience.
In the next decade, the new T5 will increase Changi Airport’s passenger handling capacity by about 50 million passengers annually.
Beyond building world-class infrastructure, we must rethink how our air hub operates.
How will passengers move from one gate to another? How will baggage be delivered across multiple terminals? How will aircraft landings and take-offs be sequenced on our runways?
These are genuinely hard problems that need both hardware and software innovations.
A new terminal alone won’t do the job. There are multiple constraints to doubling the flights that we serve.
Among other challenges, we will need a next-generation Air Traffic Management system that prioritises both safety and volume.
This is just one of the many exciting opportunities in aviation that AI can help with.
Our maritime hub at Tuas Port is another good example. It is the world's largest automated container terminal along one of world’s busiest waterways.
There are rich datasets of complex operations useful for developing exciting new solutions.
Tomorrow, my colleague, the Transport Minister, Mr Jeffrey Siow, will share more of his ambitions with you.
Our connectivity challenges clearly demonstrate where Singapore can be a living lab for the world.
Likewise in Advanced Manufacturing, the scope and scale for impact is significant.
Developments in physical and embodied AI have great relevance to our manufacturing companies.
They already operate at the cutting edge of efficiency to thrive against intense competition.
Industrial robot density in SG is about 5X the global average, and consistently one of the highest in the world.
Physical AI can help with simulations for process redesign.
Better digital twins can improve predictive maintenance, reduce material wastage and production downtime.
But it is not so simple. Ideas developed in research labs do not always perform well enough in real factory settings.
To see real impact, we will need collaborators from the hardware, software, and operational domains.
This is why we welcome NVIDIA’s new Research Lab in Singapore, focussing on embodied AI and efficient AI.
The lab will work hand-in-hand with our universities, industry partners and government agencies.
It strengthens NVIDIA’s existing foundation in Singapore, deepens our AI talent ecosystem, and unlocks new sources of innovation.
We are also developing Punggol Digital District as a frontier testbed. Years ago, not so young Singaporeans like me will remember that Punggol was a sleepy village.
But in today’s Punggol, we will create an integrated data platform, design real-world test scenarios, and rules that enable robots to safely operate across the district.
Robots can help our workers enhance service delivery to areas that are currently underserved.
A growing network of industry partners is using Punggol’s ecosystem for testing and experimentation.
These sandboxes and collaborations help to spread acceptance and adoption.
On its own, Singapore’s domestic market may not warrant so much attention.
What makes us compelling is the global network we are connected to and our track record for trusted technology adoption.
Leading companies like NVIDIA choose to deepen their presence here because we add something valuable to their global footprint.
We look forward to more of you joining them.
You will have noticed a pattern across what I shared today. The testbed at Punggol. Our ambitions for Changi T5. Etc.
In all these efforts, Singapore entities will not be passive observers, waiting for solutions to be served on a platter.
We will engage actively, address pain points along with our partners, strengthen our ecosystem to create new value for business and industries.
These partnerships are one of two critical enablers to activate our National AI Missions.
The other is governance – which we take seriously, because what we use in Singapore must be worthy of trust.
Although AI has developed at a breathless pace, governance approaches are far from settled.
But the stakes are getting higher, as AI becomes more deeply embedded in areas affecting people’s lives – healthcare, finance, transport, public services.
The risks from autonomous agents, from cybersecurity threats, from the erosion of trust in information – they do not respect national boundaries.
The benefits of AI may also become narrowly distributed. Reduced access and lack of inclusion will disadvantage smaller countries like Singapore.
Internationally-recognised rules and standards will be important. But they will take time to form.
Meanwhile, we have started to spread and shape expectations, monitor how they shape ground realities, and adjusting them as we go.
One example is the Model Governance Framework for Agentic AI, which was developed with industry inputs and launched at the World Economic Forum in January.
It has been just a few months, but we are updating the framework today, with case studies of real-world agentic deployments by companies like PwC and Workday.
We also released a case study on OpenClaw, highlighting best practices for users.
Within the government, we apply the same iterative process to our own use of agents, experimenting through sandboxes with partners like Google, and learning how to make them safe.
In parallel, we interact closely with the global community of AI safety practitioners.
In fact, on Monday, an eminent group including Professors Yoshua Bengio, Stuart Russell, Dawn Song, Zhang Ya-Qin and Max Tegmark have just updated the Singapore consensus on AI Safety Research Priorities.
These efforts recognise that much work and international cooperation is needed before we can be assured of safety-by-design, where AI is deployed responsibly in areas that matter.
Singapore sees them as an essential investment of time and treasure, which we will commit to the long term so that AI can be trusted.
The efforts I outlined are part of AI becoming mainstream in our economy and society.
As a result, the AI hub in Singapore is increasingly vibrant.
Our National AI Missions provide strong reason for leading AI firms to anchor themselves here, to develop, test and scale AI solutions that are trusted and globally relevant.
At the same time, in areas like healthcare and education, we are harnessing AI for the Public Good and to improve athe well-being of our citizens.
We will speak more about them at suitable platforms.
In conclusion, we hope your time at ATx will give you a clearer sense of what it means to build with Singapore.
I invite you to join us to make a greater positive impact.
And I wish you all a very fruitful ATx.