UK government must fix its data foundations to lead in AI, experts warn
The UK government's ambition to become a global AI superpower will stall unless it closes a significant gap in data analysis capability, confronts bias in its datasets, and rebuilds public trust in how it handles information.
A knowledge gap exists between what the commercial sector considers basic data literacy and current practice in government, according to Fawad Qureshi, field CTO at Snowflake. "What we assume is common sense or straightforward is not always so common in government," he said.
That gap matters because data underpins policy decisions, public services, and public trust. Without addressing it first, AI investments will be built on unstable ground.
Bias starts upstream, not in the model
The myth that data is neutral persists in government. It isn't. Data captures human behaviour - including human bias - and AI systems amplify those biases unless departments act deliberately to prevent it.
Investment must shift upstream, before any AI model is built. Departments need to prioritise data quality, provenance, and diversity. Context matters too.
Refugee data collection offers a concrete example. Someone fleeing across borders may view government as an adversary. If data is collected in a context of fear, it won't reflect reality. The result: biased datasets that lead to biased policy.
"If the UK wants to lead in AI, it must first lead in understanding its own data," Qureshi said.
Trust is fragile and easily broken
Public attitudes research consistently shows that trust in government data handling depends on three things: transparency, accountability, and clear public benefit.
The NHS remains one of the UK's most trusted institutions for handling data. Trust drops significantly where people perceive weaker oversight or commercial involvement, according to government surveys.
Recent Department for Science, Innovation and Technology research into a proposed National Data Library found that citizens worry about "undisclosed intentions" unless governance and transparency are clear.
Once trust breaks, it's difficult to rebuild. "When you collect data for one purpose, use it for that purpose. Don't quietly use it for something else. You will break trust - and once it's gone, it's gone," Qureshi warned.
Accountability cannot be automated
AI promises efficiency at scale. But scaling decisions is not the same as scaling accountability. Decisions made by automated systems still need to be explained, challenged, and owned by someone.
There's also a perception problem. People tolerate human mistakes more readily than machine mistakes. One failure in an automated system becomes a national story.
Synthetic data and deepfakes are reshaping the information landscape. Government needs to invest in provenance and trust frameworks alongside analytics, ensuring data can be traced, validated, and verified.
The leadership problem underneath
These issues point to a deeper problem: government lacks the talent and capability to understand what it's trying to regulate.
"You cannot regulate what you do not understand," Qureshi said. The current government model - constrained pay, heavy reliance on contractors, and structural inefficiencies - makes it difficult to attract the expertise needed.
When the best people work in the private sector, government ends up outsourcing policy work it cannot fully understand. Global competition for AI talent is intensifying. "AI is the new moon race," Qureshi said. "If you want to attract the best talent, you need the right environment - not just strategy documents."
What needs to change
Becoming an AI leader requires a systemic shift across government:
- Treat data as a strategic asset, not an afterthought
- Protect and nurture public trust as a core capability
- Design for long-term societal impact, not short-term compliance
- Embed accountability into every automated system
- Build verification into the fabric of digital services
- Create an environment where top talent wants to work in government
Qureshi summarised the stakes plainly: "Trust is earned in drops and lost in buckets. One misuse of data, and people remember it for decades."
For AI for Government professionals, the message is clear. Technical capability alone won't solve this. Leadership, governance, and people are the real constraints.
Your membership also unlocks:
