The UK has no shortage of AI strategies, summits and action plans. But a stack of policy papers doesn't retrain a civil servant, fix a legacy data system, or scale a pilot beyond a single department. The real challenge is execution.
In this episode of Policy Fix, Nesta's podcast on the ideas and reforms that matter, host Joe Owen is joined by two people who know what it actually takes to get things done inside government: James Kuht, former member of the Prime Minister's Data Science Team at No.10 and now CEO of AI workforce company Pair, and Mallory Durran, who led the government's Incubator for AI and now heads Nesta's Applied Research and Methods team.
Together they argue that the UK is at a genuine inflection point but risks squandering it by mistaking activity for progress. They explore what separates real transformation from efficiency theatre: the internal barriers of skills, culture and creaking infrastructure; the pressures of Big Tech dependency and procurement inertia; and the question of whether the public sector has the risk appetite to actually innovate.
The UK does have real strategic advantages. This episode makes the case for what they are and how to build on them. But it ends with an honest question: are the reforms being discussed genuinely equal to the scale of the opportunity?
You will come away with actionable insights on bridging the implementation gap, building national AI capability beyond running pilots, and understanding the concrete reforms needed to turn policy ambition into operational reality.
What we cover:
- Why AI success in government comes down to senior leadership clarity, not technology
- The internal barriers blocking scale: skills, culture and legacy infrastructure
- How to manage AI risk by investing in workforce experimentation
- Whether the UK can compete with Big Tech or is inevitably dependent on it
- The bold bets that could genuinely transform health and education outcomes
- What departments need to switch off to make space for AI to deliver
Guests:
James Kuht, CEO, Pair | Former Prime Minister's Data Science Team, No.10 Downing Street
Mallory Durran, Head of Applied Research and Methods, Nesta | Former Director, Government Incubator for AI
Chapters:
00:00 "The magic can really start to happen"
01:14 Introduction
02:38 Where does AI sit on the hype spectrum?
05:07 How big is the prize for public services?
07:27 AI in action: planning data and the Extract tool
11:51 The biggest barriers to AI adoption in government
15:01 Why government needs tech-native leaders
17:25 Real risk vs perceived risk in the public sector
20:37 Experimentation as a risk management strategy
23:31 What breaks when you try to scale an AI pilot?
25:46 The funding problem: how government scales tech
28:06 Can government compete for AI talent?
32:11 Insource or outsource? Getting AI capability right
34:01 What would prove the government is serious about AI?
37:04 The unglamorous reforms that will actually matter
39:22 Practical advice: where to start with AI tomorrow
42:14 Outro
