Think Data for Government 2026: Government’s data sharing challenge is now cultural rather than technical

Despite years of investment, legislative change and renewed focus on data, sharing information across government remains slow, complex and difficult.

That was the challenge posed during a panel discussion at Think Data for Government 2026 (pictured), where public sector data leaders explored why data sharing agreements still take months – or even years – to negotiate, despite widespread recognition that better data sharing is essential to improving public services.

Bringing together representatives from the Department for Work and Pensions (DWP), the Department for Transport (DfT), the UK Health Security Agency (UKHSA) and consultancy Transform, the discussion repeatedly returned to one conclusion: technology is no longer the biggest obstacle.

‘At the heart of every data problem, there is a people problem’

Matt Thompson, head of data strategy at UKHSA, said his organisation exchanges information constantly with government departments, local authorities, international partners, academia and the life sciences sector, making effective data sharing fundamental to its work.

While technology plays an important role, he argued that organisational behaviour remains the biggest challenge.

“At the heart of every data problem, there is a people problem,” he said. “We can as a department in isolation do the technological bit… it’s so much harder… to get people to understand that we need to share the risk, or to accept the risk for the benefit of someone else.”

Changing that mindset, he suggested, will take time.

Data sharing only works when everyone sees the benefit

Katherine Williamson, chief data officer at the Department for Transport, described transport as a highly fragmented sector, with information spread across operators, local authorities and national organisations.

She said successful data sharing depends on what she called the “three Ds”: discover, discuss and document.

The first challenge is simply finding the data. The second – and most important – is ensuring both organisations understand why the information is needed.

“One of the reasons a lot of data sharing fails is that both parties involved don’t necessarily have a shared understanding of why that data needs to be shared and where the value is,” she said. “Often the value is for the person that wants the data, not the person who will be sharing it.”

Only once those conversations have happened should organisations move on to documenting formal sharing agreements, she said.

‘No one wants to do the data sharing’

Diana McAuley, head of data strategy at DWP, argued that government often misunderstands what people mean when they call for better data sharing.

“The appetite for data sharing… quite often what people mean by that is data acquisition,” she said. “No one wants to do the data sharing; people want to be getting the data.”

She explained that one reason agreements can take so long is because the department supplying the information carries most of the legal and reputational risk, while the department requesting it often receives most of the benefit.

“The department that holds the data… isn’t carrying the imperative for the data that really understands the need,” she said, arguing that departments need a better shared understanding of both risk and value.

Learning from Covid

Asked how government could change attitudes towards risk, McAuley pointed to the Covid-19 pandemic as an example of data sharing happening at pace.

“I think when we think about when we’ve successfully shared data really quickly, probably Covid is the example that springs to mind,” she said.

That happened because the consequences of not sharing data were greater than the risks associated with sharing it.

“I think that’s the key… to have conversations about the risk of not sharing the data. What is the impact on citizens if we don’t share this data?”

She also suggested technology can reduce friction by automating elements of governance, pointing to DWP’s data marketplace, which can automatically generate data sharing agreements rather than forcing teams to start from a blank document every time.

Williamson added that better metadata management and standardised documentation could also simplify the process, while initiatives such as the proposed National Data Library could help create more consistent approaches across government.

Building trust

Will Lowe, chief data and AI officer at Transform, argued that public attitudes towards data sharing are closely linked to whether people understand the benefits.

Drawing comparisons with services such as Tesco Clubcard and Uber, he said consumers are often willing to share personal information when there is a clear value exchange.

Government faces a similar challenge, he suggested, particularly when dealing with highly sensitive information.

He cited work on the Department for Education’s Single Pupil View as an example of the opportunities – and complexity – of linking sensitive datasets to improve services.

Start small

Responding to audience questions about experimentation and innovation, Williamson suggested organisations should avoid trying to negotiate large-scale data sharing arrangements from the outset.

Instead, she recommended bringing data owners together to work collaboratively on small pilot projects.

“You don’t know what you want… but you can start to see what’s valuable,” she said. “The value of getting people in a room just for a day… I can’t overstate how important that really is.”

McAuley added that government should increasingly look for ways to avoid moving data altogether where possible, reducing risk while still enabling collaboration.

The discussion ended with broad agreement that improving data sharing will require more than better platforms or legislation.

Instead, progress will depend on building trust between organisations, creating shared incentives, and ensuring governance enables responsible collaboration rather than becoming a barrier to it.