The UK government could deliver billions in productivity savings if it uses artificial intelligence (AI) to help transform public services.
That’s just one of the conclusions from a National Audit Office (NAO) report published earlier this year which examined how effectively the government could maximise the opportunities — and mitigate the risks — of using AI to provide public services.
The report, Use of Artificial Intelligence in Government, examined how AI can be used in the public sector to improve internal processes, support operational decision-making, and provide services, among other things.
It revealed that while 70 percent of government bodies are either currently piloting — or planning to use — AI to carry out tasks, more needs to be done to identify the most promising use cases and learn how the technology can be implemented effectively.
Current AI implementations in government
For instance, the report revealed how HM Revenue & Customs is using AI to automatically help customers complete tasks or to aid them in finding information. In another part of the Whitehall machinery, HM Land Registry has developed an AI tool to help eliminate time-consuming manual checks by staff looking for differences between application forms and other registration documents.
Meanwhile, Natural England is using AI to create a habitat map of England called ‘Living England’ to help inform environmental policymaking. It uses machine learning and satellite images, field data records and other geospatial data to predict habitats without the need to survey the whole country.
Despite these examples, and the evidence of the NAO’s survey, the report found that AI adoption within government is still at an early stage and lacks direction regarding ownership, accountability and funding.
The NAO went on to warn that achieving wide-scale benefits will “require not just [the] adoption of new technology but also significant changes in business processes and corresponding workforce changes.”
It concluded that gains can only be realised “if government ensures its overall programme for AI adoption is ambitious and supported by a realistic plan to tackle its ageing IT systems, addresses skills gaps and improves data quality,” it said.
Challenges hindering AI adoption are nothing new
I couldn’t agree more. By mentioning legacy systems, skills shortages and data quality, the NAO has rightly zeroed in on some of the structural obstacles facing the development of AI within government.
While there has been plenty written about what AI could do, these narratives often fail to comprehend the enormity of the task that lies ahead. If IT departments could get the funding, if the problems of legacy systems and siloed data could be solved, and if the skills gap could be narrowed so that we have sufficient people to support such projects, only then could the government start to plan for AI-enhanced services. That’s a lot of ‘ifs’.
And yet, these are the structural problems that must be addressed before we start thinking about new applications and outcomes. These issues are nothing new. They’ve been around for decades. It’s just that the recent advances in AI have thrown them into sharper focus.
If we’re to see the benefits of AI in terms of streamlining services and maximising the impact of government spending then, as the NAO points out, the government needs to lead from the front. And that means starting from the ground up.
For instance, AI could be used today to ease the burden on IT teams who look after the government’s complex systems and networks. It can also be used to automate routine tasks like password resets and network troubleshooting.
Using AI to improve IT systems would help free up skilled resource
If departments were to make AI assistants an official part of their IT frameworks, the technology could bring order to the chaos of a decentralised IT environment. For hard-pressed IT pros, it would be like having an extra set of hands, but ones that never tire or falter in their accuracy. Not only would this save time, but it would also free up employees to concentrate on other high-value tasks.
Even this is not without its problems. As has already been mentioned, the pressure on the public purse is constant and means governments have to make tough decisions.
AI has to be used in the right way for government. To avoid it would leave the digital public service falling behind. The question is how to incorporate the beneficial aspects without the potential risks that AI poses. It’s important, therefore, that the government builds an ethical framework with robust data governance policies. It must also be transparent with the public about what they task AI to do.
Looking ahead, there’s little doubt that artificial intelligence has the potential to make a significant contribution to the way the civil service helps conduct government business and deliver public services. But realising its full potential hinges on addressing deep-rooted challenges like legacy systems, skills gaps, and funding constraints. It’s a big ask. It will be interesting to see how the UK’s new administration tackles these challenges.
