AI innovation should be measured by its ability to impact lives: Josephine Teo

Rather than focusing on the capabilities of artificial intelligence (AI), Singapore's main focus when deploying AI is its ability to impact lives, said Minister for Digital Development and Information Josephine Teo.

Speaking at the Digital@UNGA event in New York on Monday (Sept 21), Teo said that there are three key factors to AI innovation — impact, trust and inclusivity.

Infrastructure is less of an immediate constraint, she said, explaining that Singapore has consistently invested in its digital capabilities.

Instead, the challenge is deploying AI for impact, so that it can augment public service officers and help serve citizens better.

According to Teo, almost 70 per cent of public officers currently use AI tools to work more effectively and improve services for citizens.

"The Government is also exploring the use of AI agents to redesign entire workflows, while ensuring that officers remain accountable for key decisions," she said.

Teo also raised multiple examples of how AI is being incorporated across public agencies, including the use of machine learning to "route feedback so that a broken swing or faulty streetlamp gets to the right agency more quickly".

"The real measure of AI innovation is not how impressive the model is. It is whether it improves people's lives," she said.

At the same time, AI can also be used for more sinister purposes, and has made it easier to execute cyberattacks and run scam operations.

As such, Teo said that managing these risks and putting safeguards in place is essential to enabling people to use AI with confidence.

"Trust does not constrain innovation. In fact, trust allows innovation to scale," she said.

Teo also emphasised the importance of building inclusive AI models.

"Many AI models have weak understanding of (Southeast Asian) languages, accents, and cultural contexts," she said, highlighting the development of SEA-LION (Southeast Asian Languages in One Network) — a family of open-source large language models designed around Southeast Asian languages and cultures, including multilingual speech and code-switching.

She said: "Countries should not all have to build frontier models. Indeed, most of us will not have the capacity to do so. But we should be supported to adapt AI to our own languages, cultures, and needs."

Teo also shared that research institute AI Singapore is working with partners like the Gates Foundation, Google, and the United Nations Development Programme on Project ATLAS, to build AI-ready data sets for underrepresented languages across the globe.

"Ultimately, the success of AI should not be measured by how powerful our machines become. It should be measured by how many people are empowered by them," she said.

The minister is on a working visit to the US from Sept 21 to 24, where she will be participating in events held alongside the UN General Assembly, and engaging with leaders in industry and academia in New York and Boston.

dana.leong@asiaone.com

新加坡数码发展及新闻部长杨莉明表示，新加坡在部署人工智能时，主要关注的是人工智能对人们生活的影响，而不是人工智能的能力。

Teo周一（9月21日）在纽约举行的Digital@UNGA活动上发表讲话时表示，人工智能创新有三个关键因素——影响力、信任和包容性。

她表示，基础设施并非迫在眉睫的制约因素，并解释说新加坡一直在投资发展其数字能力。

相反，真正的挑战在于如何有效部署人工智能，使其能够增强公共服务人员的能力，更好地为公民服务。

据 Teo 称，目前近 70% 的公务员使用人工智能工具来提高工作效率并改善对公民的服务。

她表示：“政府也在探索利用人工智能代理来重新设计整个工作流程，同时确保官员对关键决策负责。”

Teo 还列举了人工智能在公共机构中得到应用的多个例子，包括使用机器学习来“路由反馈，以便更快地将损坏的秋千或故障的路灯送到正确的机构”。

“衡量人工智能创新的真正标准不在于模型有多么令人印象深刻，而在于它是否能改善人们的生活，”她说。

与此同时，人工智能也可能被用于更险恶的目的，使网络攻击和诈骗活动更容易实施。

因此，Teo表示，管理这些风险并采取保障措施对于使人们能够充满信心地使用人工智能至关重要。

“信任并不会限制创新。事实上，信任能够促进创新规模化发展，”她说道。

Teo还强调了构建包容性人工智能模型的重要性。

“许多人工智能模型对（东南亚）语言、口音和文化背景的理解很弱，”她说道，并重点介绍了 SEA-LION（东南亚语言一体化网络）的开发——这是一系列围绕东南亚语言和文化设计的开源大型语言模型，包括多语言语音和语码转换。

她说：“各国不必都去构建前沿模型。事实上，我们大多数人都没有能力这样做。但我们应该得到支持，以便将人工智能应用于我们自身的语言、文化和需求。”

Teo 还透露，新加坡人工智能研究所 (AI Singapore) 正在与盖茨基金会、谷歌和联合国开发计划署等合作伙伴共同开展 ATLAS 项目，为全球代表性不足的语言构建人工智能就绪的数据集。

“归根结底，人工智能的成功不应该以机器变得多么强大来衡量，而应该以有多少人因此受益来衡量，”她说道。

这位部长将于9月21日至24日对美国进行工作访问，期间她将参加联合国大会期间举行的活动，并与纽约和波士顿的工业界和学术界领导人进行交流。

dana.leong@asiaone.com