In our series on improving public services, we’ve explored the need for government departments to be more joined-up and outcome-focused. This second article focuses on how Artificial Intelligence (AI) can be harnessed for public good.
AI isn’t some distant technology - it’s here, reshaping societies and economies. The UK is positioning itself as a global AI leader, embedding it not just in tech innovation but deeply within public services. Amid tighter budgets and rising demand, AI is no longer optional - it’s essential.
The UK government’s Spending Review 2025 commits a substantial £2 billion investment in AI to fast-track the AI Opportunities Action Plan. This funding is a key part of a wider strategy to modernise public services and boost government productivity through digital transformation. By investing in AI technologies, the government aims to enhance efficiency and service delivery across sectors such as healthcare, education, and public administration, driving innovation, economic growth, and better outcomes for people across the UK.
A recent co-pilot trial demonstrated a time saving of 23 minutes per day for civil servants. While this sounds significant, it merely scratches the surface of AI’s true potential1. Used thoughtfully, AI can transform how public services are delivered. If treated as just a shortcut to reduce costs, we risk losing sight of the core mission: helping people. Public services aren’t fixed products; they are living ecosystems of frontline experts, delivery teams, suppliers, and communities working together toward better outcomes. AI should be an enabler in this ecosystem, not a replacement.
It’s not just about faster or cheaper services - it’s about making services more adaptive, responsive, and human-centred. The best technology is useless if it doesn’t support people in moments that matter. Many public services depend on trust, empathy, and human judgment, which AI must complement, not replace.
To unlock AI’s value, start with the outcomes you want to achieve. The most effective public services are those where everyone - frontline workers, policymakers, delivery teams and suppliers - is united by a shared purpose.
Alignment isn’t optional; it’s essential for meaningful innovation. AI alone won’t transform public services. It’s what we do with AI combined with frontline expertise that sparks real change. Innovation happens when lived experience meets technology. For example, a probation officer understands the barriers that prevent people from turning their lives around. Paired with AI’s ability to identify patterns or predict risks, this knowledge can unlock smarter interventions and better outcomes.
Failing to connect frontline insight with AI development risks building irrelevant tech and missing chances to solve problems in genuinely new ways.
Government must create feedback loops between delivery and technology teams, empowering those closest to the problem to shape solutions. AI shouldn’t just speed up today’s services - it should help us imagine and build tomorrow’s.
AI raises real fears about job security, particularly among frontline workers. Earlier this year, government leaders signalled that AI could replace some civil service roles, sparking a national debate on the future of automation in the public sector. This reflects broader efforts to save costs and drive efficiency as part of the ongoing digital revolution in government3.
With the Spending Review emphasising “do more with less,” AI offers a way to drive efficiencies. But it’s about more than cutting costs. Smart use of AI can remove repetitive, low-value tasks, freeing skilled professionals to focus on areas where human judgment truly matters. Take reducing the Home Office asylum backlog, a key government priority. Automating routine processing like document checks or triaging straightforward cases could unlock major efficiency gains. This would let caseworkers focus on complex cases where empathy and nuance are essential.
The opportunity isn’t replacing expertise but directing it where it matters most. Used solely as a cost-cutting tool, AI’s potential is wasted. But as a catalyst for smarter, more human-centred services, AI can improve both efficiency and outcomes of public services.
Public trust in AI is fragile - and rightly so. While many support using AI to cut admin and reduce delays, scepticism rises sharply when AI is involved in high-stakes decisions like sentencing or immigration.
Building that trust means earning it through transparency. We must be clear about how AI is used, what data it processes, and where humans retain final decision-making authority. Much of the AI deployed in public services is actually quite mundane- speeding up paperwork, flagging trends- but unless this is clearly communicated, suspicion grows, and with good reason.
The stakes are high. If we get trust wrong, public confidence in government digital services could be seriously damaged. But get it right, and AI becomes a trusted, visible partner supporting better outcomes.
This approach builds directly on the principles set out in the UK’s AI Opportunities Action Plan (2025), which stresses ethical AI use, transparency, and public engagement as essential to scaling AI responsibly in the public sector. Departments must go beyond transparency alone - proactively engaging the public and frontline workers by explaining, listening, and involving them is essential. Trust isn’t an afterthought; it’s earned through openness, accountability, and a commitment to serving people, not replacing them.
AI is only as good as the data behind it. Unfortunately, much government data is siloed, incomplete, or biased. Public services are messy by nature, and building AI tools on idealised datasets leads to flawed results.
This isn’t just technical - it’s a fairness issue. Poor data risks biased or inaccurate outcomes, especially affecting vulnerable people. The cost of error is real: it impacts lives, families, and communities.
Operational teams closest to services must lead in reviewing and improving data before AI deployment. They understand where the gaps and messiness lie - knowledge that can’t be captured on paper. Synthetic data can help train models safely, but only if it reflects real-world complexity. Otherwise, you’re training AI on a fantasy.
Getting data right isn’t glamorous, but it’s the foundation for fair, accurate, and effective AI. Ignoring it risks inefficiency, lost trust, and unfair outcomes.
AI shouldn’t be assessed by cost savings alone. Leaders need a fresh lens centred on people, purpose, and public value.
Ask:
AI’s true power isn’t cutting people, it’s cutting friction. The goal isn’t shrinking workforces but unlocking expertise, reducing frustration, and helping services adapt to society’s evolving needs.
This is a leadership challenge requiring bold choices, accountability, and a commitment to designing services as living, evolving ecosystems driven by outcomes, shaped by frontline insight, and built with humanity at their core.
Done right, AI can help build a public sector that’s not just efficient but profoundly more human.