Letter: Innovation gets the headlines but good institutions make them last

Dear reader, 

Governments can deploy new technology almost overnight, but changing the institution that puts it to work takes much longer. 

That gap stood out to me in our top stories for September. 

Across defence, healthcare, and digital government, what interested me wasn't simply what the organisations were introducing, but what surrounding the technology had to change for it to work well and keep working. 

The Singapore Navy is confronting this challenge as it adopts more data, AI and autonomous systems. 

Beyond the tech itself, the Navy is changing how data is classified and accessed, how people are trained, as well as their expectations for their teams as more decisions become data-driven. 

Some roles changed altogether. For example, radio operators are being retrained as cyber and network engineers as communications work shifted towards managing and securing digital networks. 

At Singapore’s Defence Science and Technology Agency (DSTA), shared cloud platforms and engineering capabilities are allowing product teams to move from experimentation to deployment much more quickly. 

A new team that once took weeks to get onto the agency’s cloud platforms can now be onboarded in as little as a day. 

This speed came from building more common structures: DSTA combined various tools and workflows, that had been previously developed separately, into shared platforms. 

Tan Tock Seng Hospital offers another lesson in what happens after a reform has already shown that it works. Its Enhanced Recovery After Surgery (ERAS) programme has delivered shorter hospital stays and fewer readmissions, and has since expanded across multiple surgical specialties. 

But leaders were careful not to describe the project as something that's simply done and dusted. Even a decade after ERAS began, teams are still auditing outcomes and training staff. 

A successful launch does not equal lasting reform. Impressive results might be attainable while enthusiasm is concentrated around a programme, but will they persist even after the novelty wears off? 

The Maldives’ Prosecutor General’s Office faced almost the opposite problem: staff were ready to use AI before the institution was ready to govern it. 

A readiness assessment found a significant gap between employees’ willingness to adopt AI and their confidence that responsible-use guidance existed. 

Rather than deploy first and work out the rules later, the Office built governance into its AI programme before using the technology on live cases, accepting a slower start so that safeguards could be established first. 

Against the backdrop of conversations about AI which so often valourise speed, it is a useful reminder that readiness sometimes means knowing when not to accelerate. 

In GovInsider's interview with Manuel Kilian, an advisor to Germany’s Federal Ministry for Digital Transformation and Government Modernisation, he said that what makes digital statecraft is a country's ability to govern well in the technological era. 

Governments frequently point to Singapore and Estonia as models of digital government. But Kilian’s argument is that the technology was only one part of what made those governments technologically robust. 

Singapore’s Singpass did not reach near-universal adoption simply because it was a good product, he says. Its success also relies on the work done to position it and weave it into other services. 

What can be seen from this month's top reads is that the same challenge of what happens after adoption takes different forms.  

With AI, the technology itself may not even stay still for an institution to organise themselves any one version of it. 

The goal then should not be to create institutions to be perfectly suited to existing technologies but to create systems of governance that equip people for when technology changes again. 

Zenith Wong 

Reporter 

GovInsider