Singapore’s digital minister talks AI agents for digital inclusion, tech sovereignty

Imagine if every citizen had an assistant to navigate digital government services.  

For the tech-savvy, this might mean convenience, but for the digitally-excluded, the assistant could be a lifeline to receive services. 

“AI agents could potentially help us enable this group of citizens to come on board the digitalisation journey,” said Singapore’s Minister for Digital Development and Information, Josephine Teo. 

She was speaking alongside government officials from Estonia and the United Arab Emirates (UAE), “From Digital to Agentic State: A New Paradigm for Governance” panel at the inaugural World Economic Forum (WEF) GovTech Day organised by the two Global Government Technology Centres in Kyiv (GGTC Kyiv) and Berlin (GGTC Berlin).

She alluded to this vision of artificial intelligence (AI) agents to support citizens as “navigators” of services. 

Minister Teo added how agentic AI could be used to provide the “human touch” to some services and ensure that technology is not a "chore” to this group of people.  

“Now, this someone can take the form of an AI agent because to organise these efforts using human agents would be very costly and difficult to do so,” she noted. 

How credible governance looks like 

To effectively transition from digital government to an agentic state, governments must first secure the foundation of reliable, trusted digital infrastructure and data. 

This was a sentiment shared by Estonia’s Minister of Economic Affairs and Industry, Erkki Keldo, and Minister Teo, in the panel.  

Minister Keldo highlighted that an interoperable digital public infrastructure (DPI) enabled Estonia to pilot several cross-border agentic AI collaborations. 

While Minister Teo acknowledged the rollout of AI agents as a “learning journey,” she underlined the importance of building the hands-on experience and credibility to regulate the technology. 

“As a government, if we ourselves are not using the agents and we do not know the difficulties of implementing these, I think we are less credible in terms of the regulations we would eventually have to put out,” she explained.  

The government needs to build public confidence, Minister Keldo added, highlighting the importance of civil servants to be skilled in applying technology and literate about the risks and how to mitigate them. 

Echoing this need for adaptation among civil servants, the UAE’s Chief of Government Services, H.E. Mohamed Bin Taliah, shared how the roles of frontline public officers are being rescoped as AI takes on more routine tasks. 

The number of government service centres has been reducing in recent years as services have become digitised, he said. 

“We took these [frontline officers] and use them as our consultants into what citizens want because they're the ones who have been engaging with the citizens for many years on a day-to-day basis understanding what they go through,” he explained. 

“This has helped us achieve a shortcut to how we improve the services,” he noted, adding that there are currently around 600 consultants who had their roles transformed to designing services within the government. 

Achieving tech sovereignty through capability 

When asked about tech sovereignty, a highly debated topic in Europe, Minister Teo explained that Singapore views the matter as one of capability over ownership of the tech stack components. 

“Sovereignty comes through by being able to exercise the necessary controls when there is a demand for it,” she specifically said. 

While governments can try to retain more sovereign control by putting in place rules of the game, they also need to have the technical expertise to assess and ensure that rules are being followed, she explained.  

She pointed to the aviation industry as an example. While Singapore does not build the planes, it leads by managing its airspace and national airline and competing based on service quality and reliability. 

What to consider in agentic AI implementation 

Minister Teo also underlined the key factors that need to be in place when governments explore the use of AI agents. 

Firstly, governments need to ensure that human officers are still present.  

A human “backstop”, as she said, must remain available for citizens who prefer personal interaction, ensuring that technology supports rather than replaces the citizen’s agency, she noted. 

Secondly, AI agents and human officers must be deployed strategically and not for technology’s sake.  

Minister Teo retains the view that the value of using AI agents is to allow the government to redirect its limited human resources towards citizens who need more personalised support, enabling a more targeted approach to digital inclusion efforts. 

Thirdly, public confidence must be assured by taking a risk-based approach to implementation. This can be achieved by deploying AI agents on low-risk tasks initially.  

“Citizens would want the assurance that you [the government] have looked at the risks. There will be certain transactions where it may not be a good idea to use AI agents. You want to start with things that people are comfortable about,” she explained. 

The panel moderator, WEF’s Head of Europe and Eurasia, Andrew Caruana Galizia, concluded the panel highlighting how an agentic state anchored in trust, clear rules and transparency could "bring governments closer to citizens, create a positive feedback loop in terms of the state’s ability to deliver [services] for its citizens and protect them from the worst excesses of advanced technologies.” 

To read more coverage of Singapore's digital government, you can click here.