Majority of UK adults unaware of how government uses AI, research finds

Three quarters of UK adults cannot name a single way the government is currently using artificial intelligence, despite growing optimism among senior public sector leaders about its potential to transform services.

A survey of 1,000 UK citizens aged 18 and over found that 75% were unable to identify any existing public sector use of AI, suggesting an awareness gap between those designing AI-enabled services and the citizens expected to benefit from them.

The 2026 UK Public Sector AI Adoption Outlook, commissioned by process automation provider Appian, also surveyed 1,000 public sector workers. It found that while 67% believe AI will improve public services over the next five years - rising to 87% among director-level leaders - only 44% of citizens share that view. Among administrative staff within the public sector, optimism falls to 40%.

Pilot projects “invisible” to the public

The disconnect may reflect how AI is currently being deployed. Nearly half (45%) of AI initiatives in government are described as bolt-on experiments or standalone tools rather than being embedded into core service workflows. As a result, many applications may be improving internal productivity without being visible to service users.

Peter Corpe, industry lead for UK public sector at Appian, said too many initiatives remain peripheral to frontline delivery: “Too much AI in the public sector is still being used as a personal productivity tool rather than embedded into the processes that actually run services,” he said. “When AI is treated as a bolt-on experiment or standalone tool, it struggles to deliver meaningful impact - our research shows nearly half of government’s application of AI falls into that trap. If organisations want AI to move beyond pilots and produce real value, it has to be integrated into core processes from the start.”

Public trust remains fragile

Public trust in government use of AI also remains limited. Fewer than half of citizens said they trust central government (39%) or local government (44%) to use AI responsibly. By comparison, retailers (60%), banks (55%) and consumer technology companies (54%) command higher levels of trust.

The NHS is a notable outlier, with a 63% net trust rating - the highest among organisations surveyed across both public and private sectors.

There is also a marked divide over AI making decisions without human oversight. While 67% of public sector workers are comfortable with AI selecting cases for tax or benefits compliance checks, only 40% of citizens agree. Similarly, 56% of officials support its use in analysing NHS scans, compared with 40% of the public.

Concerns extend beyond individual use cases. Majorities of citizens cite risks around data security and privacy (67%), job losses (63%), auditability of decisions (61%) and ethical oversight and bias (59%).

Strategy-delivery gap

Inside government, enthusiasm for AI is tempered by delivery challenges. Fewer than a third (29%) of public sector workers say their organisation is delivering on most of its AI commitments. More than a quarter (27%) report that progress is slower than planned, while 25% see a significant gap between AI strategy and implementation.

Nearly nine in ten (89%) say their organisation is not yet fully able to leverage AI.

Both officials and citizens broadly agree that improving underlying processes should take precedence over scaling new AI tools. A majority of public sector workers (55%) and citizens (56%) say existing systems and workflows must be fixed before introducing AI at scale.

Corpe added: “AI is only as good as the work you give it. This research shows strong belief in AI’s potential, but also a clear warning: without fixing the underlying processes first, AI will struggle to deliver on its promise. Serious AI is not about experimentation or standalone tools - it’s about applying intelligence to the core processes that keep public services running.”

Access the full research report here.