Leaders gather at the AI Summit in Seoul

Leaders gather at the AI Summit in Seoul, May 2024 (photo from The Guardian)

Opportunities and Challenges: The Adoption of AI in the UK Public Sector

 

Written by Ben Yardley, Learning Coordinator

 

Tech enthusiasts, investors and industry experts gather at London’s Tobacco Dock this week to network, and show-case new technologies aimed at the business community. This event follows the second AI Summit in Seoul last month (jointly hosted by the UK and the Republic of Korea) and the much-celebrated summit at Bletchley Park in November 2023.

 

These summits clearly demonstrate the UK’s aspiration to become a global leader in the adoption of artificial intelligence, as the international community seeks to rapidly explore the new opportunities artificial intelligence provides, while also mitigating its potential dangers.

 

One such opportunity is the use of AI in the delivery of public services. Prime Minister Rishi Sunak, and especially Deputy PM Oliver Dowden, have championed artificial intelligence as a way to improve outcomes and reduce costs in the UK public sector.

 

In a social media post in March, Oliver Dowden called artificial intelligence ‘genuinely a silver bullet’, continuing that, ‘it is only by the rapid adoption of AI that we will drive the savings in government needed to put us on a path to a smaller state’. With Civil Service numbers reaching 500,000 in December, key figures in government hope that artificial intelligence could be tool that helps the government meet its ambitious target to reduce the number of Civil Servants to pre-pandemic levels. An early estimate by the Central Digital and Data Office has suggested that as much as a third of tasks in the Civil Service could be automated, although this estimate did not fully examine how much this would cost or how realistically this could be carried out.

 

Although the will to use AI to transform the productivity of both the Civil Service and the broader public sector is there, significant practical and ethical constraints remain.

 

Overall strategy and acceptable use

 

Keen to realise the potential of new technologies in the public sector, the government has made progress in setting out how AI could be used in this area. The 2021 National AI Strategy details the government’s plans to make Britain an “AI Superpower” by encouraging UK investment in technology, regulation, and making sure that the public benefits from the government’s use of AI.

 

However, a March report by the National Audit Office suggested that the government still needed to develop ‘a coherent plan to support adoption of AI in the public sector’, noting that it was unclear whether responsibility for driving this initiative lay with the Cabinet Office or the Department for Science, Industry and Technology.

 

The Civil Service’s difficulty in hiring and retaining tech specialists is another obstacle; despite changes to the pay scale for Data Digital and Technology roles, there are still an estimated 4000 vacancies across government. This can lead to an overdependency on external consultants, which can be expensive and does not contribute to building institutional memory.

 

Moreover, there are still inconsistencies over exactly when and how Civil Servants should use artificial intelligence in the workplace. The government took steps to clarify Civil Servant’s use of generative AI, producing a guide in 2019 (since updated in June 2023) which sets out for which tasks Civil Servants should and should not employ artificial intelligence.

 

However, in February this year the Department for Work and Pensions announced it was issuing a blanket ban on using public AI, notably ChatGPT, for work-related matters. This policy is far stricter than the general Civil Service acceptable use guidelines. There is a risk that a lack of clarity in this area could lead to different departments taking very different approaches, worsening the longstanding issue of “departmental siloes” in Whitehall. An alternative in-house tool for the Civil Service, by the name of Redbox Copilot, is still in development, though it has recently progressed to trial stage.

 

AI and Public Ethics

 

Much of the media attention devoted to AI ethics has focused on existential risk, weaponisation or large-scale electoral interference. But there are also significant challenges to make sure that the use of artificial intelligence is compatible with public sector ethics. The issue is not only whether functions of government can be performed by AI, but whether its use is appropriate and safe.

 

In March, it was revealed that the West Midlands Police had trialled an AI system called AMY101 to screen non-emergency calls after an ethics committee report was mistakenly leaked online. Although the scheme attracted much negative attention, in particular over concerns raised in the meeting about whether “Amy” would be able to understand a Birmingham accent, it is a possibility that AI voice assistants could be used more widely in the future to process non-emergency - and maybe even emergency - calls.

 

The example of “Amy” raises difficult questions about when it is appropriate to use artificial intelligence in public services; doubtless, many members of the public would prefer to speak to a person when making a call to the emergency services.

 

The question of which public services should be automated is certainly not new, and nor is it confined to artificial intelligence. The Horizon scandal, for instance, is an example of what can happen when technology is relied upon without proper oversight, scrutiny, or lines of accountability. However, the more that the work of government becomes automated, the more pressing these issues become – especially given that some of the proposed use-cases for AI, such as processing the claims of asylum seekers, could disproportionately impact vulnerable people.

 

Oliver Dowden has announced that each department would need a minister responsible for adoption of AI under their department. They, and other key decision-makers, will have to make many difficult calculations about where artificial intelligence can be deployed, often under significant pressure to reduce costs. Whatever the result of the election in July, it will be the responsibility of the incoming government to make sure that efficiencies in the public sector do not come at the expense of accountability or high standards of public service.



Training Catalogue

Complete Training Catalogue
Jan - Dec 2024

New Course

The training aims to help civil servants in ALB sponsorship/partnership roles to gain a deeper insight into the governance and financial management arrangements with which ALBs are expected to comply.

Tailored in-house training

Civil Service College can deliver In-House training within your organisation that is exactly tailored to meet your individual training requirements.

Training Alerts

For more information on how we protect and store your submitted data, please see visit our privacy policy.