Gov.uk: Are we ready for an AI Revolution

13 January 2025

Kier Starmer has unveiled an ambitious blueprint to turbocharge AI development across the nation, promising significant growth, efficiency and innovation.

While it’s positive that the government recognises the potential of AI and is investing in its development in the UK, there’s a fundamental issue: most organisations, especially in the public sector, aren’t ready to harness this potential.

Multiple hands on ever-expanding computer keyboards to help illustrate the concept of AI

For AI to deliver real and lasting benefits, the state itself needs to work differently (or to put it another way: new tech + old state = same disappointment). Without addressing foundational challenges in clunky systems, processes and workforce preparedness, far from kickstarting the next industrial revolution, AI risks becoming just another overhyped technology that promises much, but delivers little, especially in critical sectors like healthcare.

Secondly, AI's promise of increased productivity is so far unproven – and assumes that organisations are already well equipped to leverage the technology (many public sector organisations, particularly in healthcare, are not). The real barriers to productivity are often not technological but cultural and process-based. Understanding user needs, designing effective systems, and fostering a culture of adaptability are just as crucial as the technology itself.

The National Health Service (NHS) is frequently cited as a prime candidate for AI innovation. From patient data management to streamlining diagnoses, AI appears to offer solutions to many of the system's long-standing issues. However, AI is not a cure-all for deep-seated structural problems. As one example, AI won’t resolve the NHS’ issues if, for instance, nurses still face delays of 15 minutes just to log into their laptops.

Implementing AI in an overstretched, under-resourced system will likely result in frustration, not progress. That’s why we need to avoid being swept up in the AI hype. After all, its potential can only be unlocked when integrated into systems that are designed to serve people.

Guardrails for AI success

While the Government’s investment in building technology capability in the UK is encouraging, the success of its strategy depends on laying the right groundwork. To ensure its potential is realised, here are several key actions to take:

–Address the Basics: Focus on fixing structural issues, bridging resource gaps and eliminating inefficiencies before rolling out AI solutions.

–Prioritise People: Emphasise user-centered design to ensure services meet the needs of those who will interact with them.

–Adopt Flexible Models: Move towards test-and-learn approaches – as highlighted recently by Pat McFadden.

–Build Capacity: Invest in training, tools and resources that will empower teams to implement AI effectively.

AI, while potentially transformative, is a double-edged sword—its benefits must be weighed against the risks of misuse. Already, individuals are grappling with the harms of misinformation and data fraud perpetuated by AI systems. Developing robust safeguards is just as crucial as delivering the productivity and innovation AI promises. Without protections in place, the risks could undermine trust and limit the technology's positive impact.

The need for adaptive approaches

To succeed with AI, the public sector must abandon rigid, outdated delivery models. This means adopting funding models that focus on progress and outcomes rather than long-term predictions. Teams need the autonomy to experiment, learn and adjust as they go. Without this shift in mindset, AI projects will be constrained by old processes, preventing them from reaching their true potential.

Yes, the Government’s ambitious AI strategy is commendable, but ambition alone is not enough. For AI to thrive, systemic issues within public sector organisations must be addressed first. To unlock AI’s true potential, it must be integrated into adaptable, user-centered systems that prioritise experimentation and learning. Without this approach, AI risks being thwarted by outdated structures, delivering disappointment rather than transformative change.