‘From passports to policymaking: where AI can boost the public sector’

SPONSORED

The public sector in the UK employs almost six million people and is responsible for 20% of the country’s economic output. The sheer size means that, in theory, the scope for benefits from AI-driven innovation is huge.

So what might that look like in practice? This is a big question, and one way to simplify it is to think about what government does – and where AI can really help.

We look at three massive government tasks that constitute the bulk of public sector work: communicating with the public, conferring rights to individual citizens, and making policy decisions.

Communicating with the public

Government needs to exchange information with the public. Indeed, it has an obligation to do so. People need official, reliable information about public services, benefits, subsidies, rules, regulations, policies and so on.

And government needs to get information from citizens. Observing how people encounter public services provides insight into public sector performance and can help to identify emerging problems. Understanding how people experience policy change, can provide evidence of public acceptance and legitimacy (or otherwise) of policy decisions.

AI can help with these tasks. LLM-based chatbots can respond to citizens’ queries and help them to find the information they need. LLMs can be used to automate public consultations, so that they can reach far more widely across the general public than they have before.

A key question is how we can encourage public organisations to think beyond what they currently do?  For the first time ever, we can use generative AI to create feedback loops between the public and government, at scale.

Conferring rights

Another massive chunk of government work involves conferring rights to individual citizens, such as a license or passport, entitlement to a benefit or service, or rights to residency or citizenship.

The dream for large parts of the public sector is to automate entire processes. In general, we see a lot of public backlash and poor design, development, and deployment practices in precisely these types of projects, where the aim is to automate everything.

What we advise government to do is to think of any administrative process as a string of micro-transactions. If you apply for a passport, for example, there is a string of micro-transactions that need to be made to get one. These are small things like validating a photo or cross-checking text data from the application.

Our study found that central government alone completes about one billion of these micro-transactions per year, of which 120 million micro-transactions have very high automation potential with AI. These are complex repetitive tasks like checking photos, extracting salient text, triaging documents. There lies a lot of the potential for AI in government.

The key question is how to get officials excited about automating the minutiae of government’s bureaucratic practices? It does not sound like the stuff of which dreams are made. It won’t make front page news – but this is where AI – at the moment – can lead to substantial efficiency gains.

AI for public policymaking

The third area of government work where AI can be a game-changer is improving policy decisions.

In the UK, more than 52,000 policy decisions underpin how over a trillion pounds of public money gets allocated each year. Improving the way in which policy decisions are made can lead to substantial benefits. Data science and AI can do that in ways in which traditional statistical techniques cannot.

We have a tremendous opportunity to adapt our economic models for an age when every individual and company generate massive amounts of data on a daily basis, where the technologies that we have allow us to develop better answers to the questions that policy-makers have been asking for years. What happens if a big company closes? Can the excess capacity be absorbed by the local economy? Where are the skill shortages and who are the workers that can most easily be retrained to fill them?

Subscribe here to our daily newsletter roundup of Labour news, analysis and comment– and follow us on Bluesky, WhatsApp, X and Facebook.

How to make it happen? Adoption is key.

These three areas of AI opportunity correspond to massive government tasks: communicating with the public, conferring rights to individual citizens, and making policy decisions. These are ripe with potential to achieve efficiency and productivity gains.

But they won’t happen by magic. The government have bold aspirations for AI, detailed in the AI opportunities action plan and other policy documents. Adoption is crucial to achieving their aims. But the question of how to drive adoption receives less attention.

AI will only lead to better public services if citizens adopt them in a widespread way. If productivity is the most hoped-for benefit of AI, inequity is the most feared harm. So it is crucial to avoid AI exacerbating existing inequalities. There is mounting survey evidence to suggest that those with low incomes perceive less benefits and hold more concerns about AI applications. So do those with low levels of digital access and skills. The government needs to ensure that their digital inclusion plan is integrated with AI plans.

AI adoption must be stimulated right across the public sector in a comprehensive way. That means integrating AI tools into public services and bureaucratic processes. Economic modelling for policy requires the generation of good local data. Public service professionals are already using AI but in a disorganized fashion without clear guidelines. Governmentwide deals with AI giants will not resolve problems caused by legacy systems that lock departments in the past.

That will mean AI capability and expertise being distributed across government. There is now a strong digital centre for government – the remodeled Government Digital Service. They will need to work in concert with departments and agencies at all tiers of government to reassure citizens that AI-powered public services are progressive, easy-to-use and trustworthy. And to help and inspire public servants to make the most of AI.

• SHARE: If you have anything to share that we should be looking into or publishing about this story – or any other topic involving Labour– contact us (strictly anonymously if you wish) at [email protected].
• SUBSCRIBE: Sign up to LabourList’s morning email here for the best briefing on everything Labour, every weekday morning.
• DONATE: If you value our work, please chip in a few pounds a week and become one of our supporters, helping sustain and expand our coverage.
• PARTNER: If you or your organisation might be interested in partnering with us on sponsored events or projects, email [email protected].
• ADVERTISE: If your organisation would like to advertise or run sponsored pieces on LabourList‘s daily newsletter or website, contact our exclusive ad partners Total Politics at [email protected].