How can government ensure public trust in data?

While trust in UK-wide institutions like politicians and the media is plummeting, confidence in official statistics remains remarkably strong. Against this backdrop, the recent Think Data for Government saw senior data leaders from government and industry (pictured) explore how data professionals can preserve and enhance that public confidence.

The panel’s focus: how to embed trustworthiness into everything from data collection, to analysis, to public-facing communications.

The pandemic: Transparency under pressure

Lucy Vickers, CDO and chief statistician at the Department of Health & Social Care (DHSC), reflected on her experience during the Covid‑19 crisis.

“My experience… really harnessed and cemented my view the importance of trust in… statistics and data, because it really thrust health and care stats and data… into the headlines.”

When Test and Trace launched in May 2020, she noted: “We stood up statistics… I’ve never worked on something where you publish something and one minute later, it’s on BBC headlines; it’s there for people to see. The use of that data enabled politicians and the public to make use of those data in the public domain.

“That’s really cemented my leaning into the importance of people believing the data. The trust that was there in terms of publishing them in a transparent, accessible way, because they were supporting the choices that people were making, people as individuals, but also businesses and the politicians were making as we were progressing. I think that today, the appetite is still there about holding to account that cornerstone of democracy.”

Defra embedding openness

Sue Bateman, CDO at the Department for Environment Food and Rural Affairs (Defra), reflected on her time at GDS, where thinking was focused on data sharing in the public good in the form of open data, ethics, and more recently algorithmic transparency.

“I was fortunate to have that bird’s eye view and get involved in some initiatives where you could see that it was quite a cultural challenge for organisations, my own teams – and myself at times – to be so open and put yourselves in a vulnerable position around scrutiny,” she said.

Bateman also emphasised practical steps Defra is taking to normalise transparency. “Over the last year… as we’ve developed a new roadmap, we’ve taken a conscious decision… engaging with partners from industry… and people in other departments… it’s more time‑consuming than going away and doing it – you’re hopefully building out and mitigating the risks.”

A CTO’s view

Fawad Qureshi, global field CTO at Snowflake, recounted how public trust in data can be lost.

“Ten years ago, I used to live in the Netherlands,” he said. “There was a famous scandal at the Dutch tax authority. A black box AI model accused 35,000 families of childcare fraud and sent them notices that they needed to immediately repay €100,000,” he said, resulting in false accusations and even with people committing suicide.

“This is what we need to think about when we are dealing in the public sector and civil services,” he said. “Lives are at stake.”

From there, Qureshi has created a five‑point framework for trust:

1. Transparency leads to trust: Your model must be explainable, reproducible, and stand up under things like ICO/GDPR scrutiny.
2. Clarity beats complexity: Data must be consumable, and understandable, and can be connected to other data that exists in the public and the commercial space.
3. Design for the public, not just policymakers: Designing inclusive data products means using plain language, offering context, and anticipating varying levels of data literacy
4. View the solution in totality: Policy making has an impact across society. Government needs to consider negative externalities, and about concepts such as national data libraries, to connect data across various agencies.
5. Consistency creates credibility: Build the entire solution in a repeatable format. There needs to be a consistent and constant stream of data.

Said Qureshi: “Trust is earned in drops, but lost in buckets.”