Solid State

The government has pledged to “mainline AI into the veins” of the nation. This ambition can only succeed if the public and workers are key partners in the process. The history of digital transformation is riddled with cautionary tales about what happens when the people affected are excluded from decisions. The National Programme for IT (NPfIT), which ran from 2002–2011, stands as a monument to such failures. A lack of engagement tanked the project at a cost of more than £10bn.

By consulting neither the people who used public services nor the workers providing them, the government practically guaranteed poor decision making, inadequate services, and a significant loss of trust that left a promising project facing mass rejection.

With public distrust in AI rippling through the UK, history is at risk of repeating itself. Only 25 per cent of citizens trust the government with their personal data and only 30 per cent trust the UK government to use AI in a safe way. Such a significant lack of trust among the electorate does not bode well for the government’s hopes of utilising AI to improve the efficiency of our public services. As the lead of a 2025 global study into trust, attitudes and use of AI puts it: “the public’s trust of AI technologies…is central to sustained acceptance and adoption”.

There have also been missed opportunities to secure buy-in from civil servants. A recent FDA survey of more than 2,000 civil servants discovered that, while the majority agree that AI tools and systems will improve how they serve the public, the implementation on the ground paints a different picture. 64 per cent stated that they had not been consulted on the use of AI in their work, with the majority of those not consulted saying that they wanted to be more involved in shaping how their departments adopt AI.

Consulting civil servants directly for their opinions and experiences also improves the likelihood of successful AI implementation. In the words of one FDA respondent, through consultation, civil servants can make “the use of AI feel more practicable, tangible and accessible…including to those who feel nervous about this change”.

The case for engaging civil servants is straightforward. Some have argued, however, that extensive consultations of the public would be in tension with the government’s ambition to roll out AI quickly. Running consultations often comes a steep cost, and analysing the responses is a time-consuming activity.

This needn’t be the case. Examples from abroad and the UK show it is possible to engage effectively and efficiently – fittingly, by harnessing the power of AI.

Over 6,000 miles away in Taiwan, the pioneering ‘cyber ambassador’ and former digital affairs minister, Audrey Tang, has shown how AI can be co-governed with the people. She introduced AI to act as a facilitator in the Taiwanese ‘alignment assembly ‘process, enabling hundreds of people to deliberate at once through real-time transcripts, minimising disruptions, and encouraging quiet participants to speak up.

In March 2024, one of these alignment assemblies took place over the course of just one afternoon, with 450 Taiwanese citizens producing regulation recommendations combatting online disinformation that would ultimately become law. This showcases how a quick online deliberation can build trust between the public and the government through the power of co-creation.

These digital democracy efforts have also begun to reach UK shores. The Waves project is currently embarking on the largest trial of AI-powered digital democracy in the UK to date. Integrating existing digital tools with new open-source technology, Waves seeks to simplify and bring down the cost of public deliberation for local government. The aim is to allow more of the public to come together to tackle contentious local issues, develop policies that work for them, and build trust in local government, starting with pilot councils in Camden and South Staffordshire.

The public, workers, and the policy making process all benefit from greater consultation through deliberation efforts. For the public, such consultation engenders trust in government systems and empowers them with more direct access to shape the policies that will most significantly affect their lives. Policymaking is improved, as AI-enabled public deliberations produce effective summaries of group discussions, insights into public opinion and sharper insights into what tools are needed ‘at the coalface’.

The work that civil servants produce, and their own working lives, can be improved by effective digital transformation. But effective transformation will not come without the kind of thoughtful consultation processes as described above. Heard early on, the views and experiences of civil servants can clear a path for a more seamless digital transition.

As another civil servant survey respondent put it: “involve us from the start”. That’s a message that the government should apply to both the public and public servants if they want AI to succeed in reforming the state.

Image credit: Mario gogh via unsplash