What the public sector AI debate is really about?

Innovation is only part of the story

Artificial intelligence is already becoming embedded in public services across Europe, helping organisations improve service delivery and enabling teams to work more efficiently. While adoption is growing, the conversation surrounding AI continues to focus on a different question: who is responsible for ensuring it is deployed safely, transparently and in the public interest?

New CARMA analysis suggests that governance has become just as important to the discussion as innovation. Regulators account for 39 percent of voices in public sector AI coverage, ahead of CEOs at 25 percent and academics at 19 percent. This tells us something important about the way AI is currently being understood. Rather than being positioned as a technology story, it is increasingly being discussed as a question of policy and accountability.

Trust is becoming the critical factor

It’s easy to see why governance is attracting so much attention when you look at the issues people are worried about.

CARMA’s analysis found that the cybercrime and AI-enabled fraud emerge most frequently, appearing in 53 percent of discussions. Misinformation and deepfakes feature in half of all coverage analysed, while concerns about systems operating without meaningful human oversight are present in 29 percent. Data integrity, security and public safety also remain recurring themes.

In many ways, public sector AI resembles the construction of a bridge. Innovation determines how ambitious the design can be, but governance provides the foundations that allow people to trust it enough to cross. Without accountability and transparency, even the most capable technology risks creating uncertainty rather than confidence.

At the same time, the appetite for AI is clearly growing. Gallup data shows that 43 percent of public sector employees now use AI at least a few times a year, up from 17 percent in 2023. More than one in five employees are already using it frequently as part of their working routines.

This is a real contrast. While headlines often focus on risks and regulation, sector professionals are already experimenting with AI tools and integrating them into everyday processes. In many organisations, adoption is happening quietly and being driven by practical needs rather than grand strategies.

From pilots to frontline services

Examples of this can already be seen at a local level. Peterborough City Council has introduced an AI assistant called “Hey Geraldine” to support social workers and care teams. Meanwhile, North Yorkshire Council has been exploring the use of AI within children’s social care, demonstrating how early pilots are beginning to move into live service environments.

These examples are important because they illustrate how AI is gradually becoming part of service delivery rather than remaining confined to proof-of-concept exercises. For local authorities facing increasing demand and workforce pressures, AI has the potential to reduce administrative burdens, support case management and free professionals to spend more time on work that requires empathy, judgement and human interaction.

However, as deployments become more widespread, expectations around governance will inevitably grow. Citizens will want assurance that systems are explainable and employees will need confidence that AI is not replacing professional expertise. While leaders will require robust frameworks to understand where accountability sits when decisions involve automated processes.

Building confidence alongside capability

The pace of investment suggests that public sector organisations recognise AI’s potential. According to Tussell, government AI procurement grew rapidly during 2025, with the number of contracts increasing by 54 percent compared with the previous year. Total spend reached £1.17 billion, more than double the figure recorded in 2024. Momentum has continued into 2026, with hundreds of millions of pounds already committed and new providers entering the market.

This highlights the role AI can play in transforming public services but also leaves public sector leaders with a balancing act. They must encourage experimentation and innovation while maintaining a level of accountability that citizens expect.

The conversation around public sector AI is often presented as a choice between moving quickly or proceeding with caution. In reality, the most successful organisations are likely to be those that recognise governance is not a barrier to innovation but an enabler of it.