The UK's digital infrastructure is not ready for a 21st-century digital government

There is a large amount of political discussion in the UK about how digital technologies may or may not be used to transform the public sector and create cost savings. However, these debates often focus on specific services or specific agencies rather than offering a holistic overview. The truth is, digitalisation in the UK is not doing all that well. I am writing this overview of the current state of digital government in the UK to inform some of these discussions better.

The UK wants to transform its public services, drive effectiveness and efficiency gains, and cut costs. How will they do this? Technology. While technology certainly can drive all of these outcomes, and as fun as it is to talk about a public service powered by AI and drones, the UK has an incredible amount of work to do.

The current situation is akin to a digital “landlord special”, where effort is devoted to making things look nice and pretty rather than addressing important problems that lie below. It would be easy to point out the service delivery failures, from recent issues at the Home Office’s immigration database to the British Post Office scandal, but that is not the point of this article.

Instead, what I hope to do, is go beyond looking at the individual failures of specific services (of which there are many) and write about what is going on behind the scenes.

Let’s start by providing some context.

On paper, things look great in the UK. In the recent OECD digital government index the UK is ranked third. However, the UK remains below the OECD average on the “data-driven public sector” dimension. This is a recurring theme. The front end may look nice, but that is where it stops. The infrastructure is antiquated, data doesn’t move, systems are not linked up or interoperable, and decision-makers cannot fully leverage the digital systems.?

None of these critiques are new, and they have been discussed for decades. Some of the more recent examples include the 2019 report on digital government that identified numerous challenges with culture, leadership, technology, and skills. In 2023, the audit report “Digital Transformation in government: addressing the barriers to efficiency” pointed out that only 4% of civil servants were digital professionals, that the existing skills gap was getting worse, that legacy infrastructure was still pervasive, and that support from senior leaders may not always exist. The report further argues that many services were not at a ‘great’ standard and that “past approaches have focussed on improving websites and front-end screens, making services ‘look good but not be good’”

The government has acknowledged these issues, as they have for decades, and in the 2022 to 2025 digital roadmap they wrote:

"The UK government still lags behind other sectors. Our services are often slow, difficult to use and expensive to deliver… Data quality is inconsistent and frequently poor and effective data sharing between departments is limited. We are held back by costly and outdated technology and we do not leverage our scale in technology procurement. We are failing to attract top digital talent or to build capability in-house at scale and our leaders are not yet as skilled in digital leadership as they need to be."

These difficulties were further reiterated by Gareth Davies (C&AG) in January of this year. He argued that efficiency gains from technology were not a given. Instead, this would require hard work to “replace antiquated IT systems, improve the quality and shareability of data, and recruit and retain scarce skills in high demand across the economy”. For some organizations (such as DEFRA) “[they] would need to spend more than three-quarters of its digital budget on addressing legacy system issues”.?

It is not possible to build an innovative, robust, and resilient digital government on top of the infrastructure built almost 10 years before the Internet existed.

Digital technology is not a cure-all or magic device that you can throw at any problem and make it go away. It is expensive and it requires maintenance. The longer this fact is ignored, the more expensive and complicated a true digital transformation will be.

The issue with legacy technology is not only one of the costs (though the UK is spending billions every year just to keep existing services running); it also has real tangible impacts.

• For healthcare, an opinion piece in the BMJ written in 2022 pointed out that “failing IT infrastructure is undermining safe healthcare” further arguing that the “digital future will not materialise without closer attention to crumbling IT infrastructure and poor user experiences”. A year later, in 2023, a BBC investigation found that IT failures were causing patient deaths.
• The police still rely on the police national computer (PNC) that was built in 1974. It is incredibly out of date and efforts to modernize it have thus far fallen short, with a recent report highlighting that its new cloud-based system has grown in costs by 68% to ￡1.1 billion. The same report further stated that it was not clear how the new system would be delivered by 2025-26.
• In the welfare sector, an investigation into the underpayment of state pensions stated that the legacy system launched in 1988 to handle pension payments led to the underpayment of “an estimated 134,000 pensioners over ￡1 billion, or an average of ￡8,900 each.”
• The most recent ransomware attack on the British Library was devastating, but made worse due to the “reliance on legacy infrastructure”.

Beyond numerous problems due to legacy digital infrastructure, there are further difficulties with data exchange and interoperability.

Sharing and using data is currently hard to do, if it is done at all. This leads to lower quality of data overall, multiple copies of data existing, poorer service delivery, and an ability to make decisions guided by strong and validated datasets. Initial efforts have been made to address this with the creation of the integrated data service (IDS), but this does not go far enough.

Failure to create, support, and implement widespread data exchange will always limit the ability of the UK’s digital government to succeed and create value. In the current discussions around AI, the implications of this are especially important.

As uncovered by the most recent audit report on AI in the UK, “large quantities of good-quality data are important to train, test and deploy AI models. Our survey found that limited access to good-quality data was a barrier to implementing AI”.

What emerges is a picture of a digital government that is failing.

The infrastructure is old, antiquated, and costly. It is not being given the support that it needs. The longer that it is left unaddressed, the higher the costs will be to fix it. Beyond costs, there are also clear cybersecurity risks, putting many organizations and the data they hold at risk. Yes, there is a large amount of interesting work being done to create new and user-friendly digital services, but these these are simply hiding the bigger problem.

A great example of this is the GDS’ recent pilot program to launch an AI tool to help users navigate Gov.UK. At first glance, this seems like a great use case of a new technology. However, what is also revealed, is that the services on offer are so complex and complicated users may need the help of an AI-based system to help them navigate it.

With little ability to share and move data between organizations and services, it is difficult to progress and build truly innovative and impactful digital services.?