Dame Antonia Romeo, the new head of the civil service, wants to reform how government works. She wants to "drive things through the system, make things happen and find solutions". And she sees artificial intelligence as critical to that end.
This ambition is exactly what the role demands, and a promising start for the government's "move fast and fix things" agenda.
Getting AI right could make public services more human. Because, by making them more effective and sustainable, the state could give people more of what they need, while asking less of them.
A recent report by the Fabian Society, supported by the FDA, argues that civil servants are well-placed to help. The report, which forms part of our technology work programme, offering solutions to make lives across the UK better because of technology says civil servants could adopt time-saving technologies that free up capacity for other value-adding activities; innovate better ways of doing things with AI; and reimagine and transform public services to make them more human.
Fortunately, many are eager to get stuck in. We surveyed 2,000 civil service managers and found that AI adoption is already widespread, at 63%. And while just 27% have used AI to improve how their team does things, a further 45% would like to do so in future.
This reflects both a frustration with the status quo, and a recognition of the potential of AI. 55% disagreed that their department uses resources as efficiently as possible. Meanwhile, 62% agreed that AI will improve how their department serves the public.
To leverage this enthusiasm, the government must resist the temptation to blame workers for the mutually frustrating pace of change over the past decade – and instead ask what they need to succeed.
Critically, rushed cuts are not the answer. It is right to have an eye on efficiency. But successful transformation entails investment in financial and human capital, and a reconfiguration of work processes, with some productivity loss during a transition, before full benefits materialise. When headcount shrinks too quickly, remaining staff become overwhelmed, services fail, and long-term improvements are deprioritised.
This is already happening in some parts of government. One survey respondent reported: “We have cut civil servants where severely needed, meaning that we are stretched so thin no-one is able to do their jobs properly.” Another said: “We have been in permanent reorganisation which stifles rather than promotes innovation.”
To realise the full potential of AI, change must be done with workers, not to them. History has shown that transformation can only succeed if social partnership is taken seriously. Workers who are effectively engaged are more likely to buy into and adopt new ways of working. Our survey found that 66% of those we surveyed want to be more involved in shaping how their department adopts AI – yet just 29% have been consulted on AI at work. Clearly, there is work to do.
Our report sets out how the government can accelerate civil service AI adoption in practice. We show how the government can develop talent and attract the best and brightest – through modular learning pathways, paid time off to learn, and by investing in leaders that are as comfortable using AI as navigating Whitehall. We urge the government to provide free AI tools and create systems to roll out what works more widely. We propose how the government can back the doers – through delivery bonuses, reinvesting efficiency savings in the workforce, a network of AI champions to help teams put AI into practice, and tackling red tape through a “Barrier Busting Taskforce”.
Perhaps most importantly, we urge the government to engage workers and trade unions and mobilise the machinery of government, to deliver a clear vision where everyone understands their role, and to make AI synonymous with good and meaningful work.
Civil servants are not ‘blockers’, standing in the way of progress. Given the right conditions, they are ‘builders’, ready and willing to innovate to improve public services. The government's digital ambitions will only be realised if it treats them that way.
Sasjkia Otto is a senior researcher at the Fabian Society and a former civil servant who worked in digital infrastructure and data directorates for the-then Department for Digital, Culture, Media and Sport. She is the co-author of Adopt, innovate, transform: How civil servants can help deliver a smarter state using AI