The British government is committed to artificial intelligence as a cornerstone of its growth strategy. Senior ministers have set out one of the world's most ambitious AI agendas. Yet new polling data suggests the public remains unconvinced that Whitehall can be trusted to pull it off.
According to fresh research commissioned by Re:State and Deloitte,51 percent of Britons worry that AI will reduce human contact in public services, 50 percent fear job losses from automation, 47 percent are concerned about over-reliance on technology reducing human oversight, and 46 percent worry about privacy and data security. Overall,37 percent see AI as a risk to public services compared with 23 percent seeing it as an opportunity.
This represents a legitimately difficult governing challenge. The economic case for AI in government is real. Administrative automation could free skilled civil servants to focus on casework and decision-making rather than paperwork. Yet the public's hesitation is not irrational. It flows from years of digital disappointment.Almost a third of respondents felt digital technology had made it worse when dealing with public services and having to repeat information; 36 percent said it made no difference.
The trust deficit runs deeper than general technophobia.Only 39 percent of citizens trust central government to use AI responsibly, a figure that trails behind both the NHS and private sector retailers. This hints at a wider institutional problem.75 percent of UK citizens were unable to identify any existing public sector use of AI, suggesting the government's AI initiatives remain invisible to most people. You cannot build confidence in technology you do not know exists.
Inside government, the picture is just as troubling.63 percent of UK civil servants report knowing "a little" or "nothing at all" about AI, and more than two in five remain unconfident in their ability to use AI tools.Only 38 percent believe AI is being used effectively within their team.
This gap between ambition and readiness has created shadow AI. When clear guidance is absent, capable staff find workarounds.Government data flows through personal accounts with no oversight.Globally, 64 percent of public servants in low-enablement environments use personal logins for AI at work, and 70 percent use AI without their manager knowing, with 2 in 5 UK civil servants unsure what they are even permitted to use AI for. This creates genuine data protection risks.
Yet public opinion is more nuanced than simple technophobia.Four in ten Britons are comfortable with AI handling administrative tasks in the public sector to save costs. On the NHS specifically, the picture is mixed.Half prefer human-led triage systems, citing trust in human judgment, while 38 percent support AI to expedite the process and reduce waiting times.
The counterargument from government supporters is straightforward: delay is itself costly.AI presents opportunities to improve public services, and the government is taking forward recommendations on adoption including a "Scan, Pilot, Scale" approach across public services. Other countries are advancing rapidly, and Britain risks falling behind in a technology race that will define economic competitiveness.
Yet this legitimate urgency collides with equally legitimate public caution.Six in ten Britons believe the government should adopt a cautious approach to AI, prioritising job protection and giving people time to adapt over rapid development. That is not backwardness. It is a reasonable demand that change be managed carefully when public services touch people's lives.
The way forward likely requires Whitehall to move in both directions at once: genuinely accelerating deployment where the evidence supports it, whilst building public understanding through transparent communication about what AI actually does in practice.In the UK, 37 percent of public servants report receiving some form of AI training, and for those who do receive it, 75 percent report that they find AI easy to use. That gap between confidence and capability is worth closing.
The risk of the current path is that government proceeds with an ambitious agenda while public and workforce confidence erode further. The risk of overcaution is that Britain cedes technological advantage to rivals whilst real opportunities to improve service delivery go unrealised. The honest answer is that reasonable people disagree on where the balance should lie. What is not debatable is that moving faster without addressing the trust and readiness gaps is a recipe for both implementation failure and political backlash.