UK deputy PM says embracing AI ‘will reduce civil service headcount’, Estonia’s ‘personal government’ vision, and more

By on 07/03/2024 | Updated on 07/03/2024
Image: Steve Johnson on Unsplash

Welcome to Global Government Forum’s AI Monitor

The use of AI could potentially transform public services, with governments examining uses that range from automating back-office activities such as finance and HR, to deploying AI to notify citizens on their entitlements.

To monitor the work of government in this important area, Global Government Forum is launching a monthly newsletter to share the latest news on how governments around the world are using AI.

We will be sharing trailblazing examples of using AI in government and exploring what public services need to do to make the most of technological advances.

We hope you enjoy this first edition, and please keep your eyes peeled for more leading analysis of the work of governments around the world in future GGF newsletters.

And please share your thoughts and news stories with [email protected]

Embracing AI ‘will reduce civil service headcount’ says deputy PM

Dowden’s AI campaign continues: The UK’s deputy prime minister Oliver Dowden has called artificial intelligence the only “sustainable path to headcount reduction” in the civil service, a plan which government intends to pursue to pay for tax cuts.

Whitehall’s silver bullet? Dowden’s comments came in a briefing with the press on 29 February, where he conceded that it was “hard to put a value” on the potential savings. However, he cited estimates from the Institute for Public Policy Research (IPPR), which suggest these could reach up to £24bn (US$30.5bn). “AI is potentially – and I don’t say this lightly – a ‘silver bullet’,” Dowden said, adding that its powers to transform the government for the better included “increased productivity, vast efficiency savings, and improved services”.

Dowden’s mission: The deputy PM has asserted the importance of integrating AI capabilities in Whitehall on several occasions over the last 12 months. In January, Dowden said that officials who feel “stifled” or “bogged down” by outdated processes could look forward to a new era in which burdensome administrative tasks are taken care of so that they can get on with more important work.

Heads no longer up: Dowden’s comments also come as his UK cabinet colleague, chancellor Jeremy Hunt, has capped the size of the UK civil service as part of an efficiency drive. Hunt has said that technology such as AI should be able to help government become more efficient, meaning the civil service can be capped at a headcount of around 488,000 – with the aim to eventually reduce its size by 66,000 to pre-pandemic levels.

Strategy to come: In the UK Budget yesterday, it was announced that the Cabinet Office will set out a “strategic vision” for how the civil service and wider public sector can take advantage of the opportunities of AI. The Cabinet Office is also working with departments to finalise AI adoption plans, which will form part of the UK government’s next Spending Review. GGF’s sister title Global Government Fintech has reported that UK’s Companies House is already increasing its data analytics and AI investment.

Union responds: Shortly after Dowden’s briefing, general secretary of the Public and Commercial Services Union (PCS), Fran Heathcote, said that though the union “recognises that AI is an inevitable development”, its interests in the potential contrasted with Dowden’s own goal of “mass job losses”. “We want AI to enhance jobs and not to degrade them,” Heathcote said in a statement on 1 March. Urging Dowden to “come clean” about his intentions, she warned that the plan to cull 66,000 civil service jobs would see “machines…in charge” of citizens’ lives across departments, from employment and pensions, to the justice system and tax.

Show us the algorithm: The UK government will soon require all central government departments to release details of algorithms used to support decisions.

The requirement follows the creation two years ago of The Algorithmic Transparency Recording Standard (ATRS). ATRS is overseen by the Central Digital and Data Office and the former Centre for Data Ethics and Innovation, now called the Responsible Technology Adoption Unit, which falls under the Department for Science, Innovation and Technology.

So far, there have been seven algorithm reports published, four of which are from central government departments. In response to a consultation on its proposals for the regulation of AI, the government said the requirement would expand to “all departments” and “across the broader public sector” over time.  

Some departments are expected to be exempt from the requirement, however. These will be revealed in the coming weeks, along with a deadline by which departments must submit their algorithm details.

UK government finds AI ‘hallucinations’ in pilots

With Dowden’s comments illustrating the government’s direction around the use of AI, a recent session at the Estonian embassy in London examined the progress of digital transformation – using AI and other tools – in Estonia and the UK.

Whitehall AI pilots: The UK government is undertaking a series of pilots of AI in the civil service, and Paul Willmott, the chair of the UK government’s Central Digital and Data Office, shared some details of how AI is being used.

From job centres to red boxes: Among the schemes that Willmott set out was the use of AI to assist civil servants in providing advice to jobseekers in job centres, as well as potentially streamlining the contents of ministerial red boxes.

Ask the chatbot: But the headline revelation from Willmott’s comments at the session at the Embassy of Estonia on the technological developments that could deliver more personalised government were around potential use of an AI chatbot on the UK government’s homepage.

Speak your mind: “The aspiration here is that, rather than having to search, you can chat to GOV.UK and find what you want, and it will give you some advice on where to go and what to do,” he said.

Ghosts in the machine: But there are problems with making sure the information is right, says Willmott. “The challenge we are having is exactly the same as in the commercial sector, which is how to deal with that 1% of hallucinations, where you’re fed in the wrong inputs, the machine starts to get misleading or imaginative.”

Resolve to evolve: Resolving this will require “help from the foundational model providers”, he said, but “until we’ve kind of managed to iron that out, we won’t be able to cut that service”.

Better red boxing: The government is also using AI to streamline ministerial red boxes. These traditional red briefcases held by UK government ministers are full of paperwork for ministers to get through. “It’s still very paper-driven, in most cases, [and] we are testing use of AI to basically absorb that material, prioritise it, support it, present the information in a more digestible way,” Willmott said. “It’s very early days, but, of course, these are high productivity costs of decision-making.”

Processing power: Additionally, the government is looking at ways to use AI to improve government processes by using pattern recognition and decision-making APIs (Application Programming Interfaces). “When you’ve got a complex decision with multiple inputs, such as whether someone qualifies for citizenship benefit, that still very much a human-driven process at the moment. A good well-trained AI should be able to make that decision for a subset of cases, and then apply that for ones that really, really need help.”

Moving the dial: Willmott said that his mantra on the use of AI in government to ministers “for years” had been that “you can’t have the cherry unless you prepared to pay for the cake”. This means that government needs to pay for the underpinning technological infrastructure through moving systems to the cloud and investing in data standardisation.

“You’re starting to get some traction about the need to really invest in technology in the right way,” he said. “We are managing to move the needle of AI through pilots and through the set of regulatory or advice frameworks that are already in the public domain.”

Read more: US government taps agencies for AI project ideas

Learn more about AI in government at Innovation 2024

Global Government Forum’s annual Innovation conference will be held on 19-20 March at the ExCeL in London. Co-hosted by the UK Government, UK Civil Service and the Cabinet Office, this unique exhibition and conference brings together government leaders from around the world to share how to make public service transformation happen.

This year’s two-day agenda will examine how governments are innovating in areas including service delivery, skills, artificial intelligence and machine learning, sustainability, and digital and data.

Register now to attend these sessions on AI at Innovation 2024

Tuesday 19 March

  • Fireside chat: Innovation in artificial intelligence and machine learning (11:10-11:40, Innovate stage)
  • Artificial Intelligence in public services: how is government using AI to improve services for citizens? (11.55-12.45, Innovate stage, supported by knowledge partner Nortal)
  • How to get the public sector ready for AI (16.10-17.00, Innovate stage, supported by knowledge partner Google Cloud)

Wednesday 20 March

  • AI in society: how governments can tackle mis and disinformation generated by artificial intelligence tools (12.05-12.55, Innovate stage)
  • RPA in the civil service – how departments can use automation to boost government efficiency (13.55-14.45, Accelerate stage)

Speakers include Yvan Gauthier, senior data scientist with the National Research Council’s Digital Technologies Division, who leads the AI Accelerator for the Government of Canada. Also billed are Weng Wanyi, who runs both Government Data and Government AI teams at the Smart Nation Group of Singapore, and DC-based Kyleigh Russ, who advises leadership within the US Office of Personnel Management on recruitment, with a special focus on technical and early career talent.

Find out more about Innovation 2024 and register to attend here.


Is ‘personal government’ the future?

During the session at the Estonian embassy in early February, Tiit Riisalo, Estonia’s minister of economic affairs and communications, set out a plan for the future of the state powered by AI and technology – personal government.

What’s that? Personal government is what Riisalo calls government for the post-digital era. Estonia anticipates becoming a 100% digitalised society by the end of 2024, and personal government is about making sure government keeps pace.

Personal government has five core characteristics to improve public services.

Human-centric: This means focusing on the citizen experience. Every government official should be a regular walk-in client of their own services to keep the focus on what the citizen should experience and then offer exactly that.

Accessible: Estonia’s government says that, in contrast to fears that e-services create a digital divide for vulnerable groups, digital services can in fact be more accessible for marginalised groups. Personal government services are intended to address that issue by making services available to all.

Proactive: Services are improved for citizens by anticipating the need or running invisible processes for the user. This means governments no longer wait for the user to initiate the consumption of the service but proactively put them into effect when the user has expressed consent.

Trustworthy: Estonia’s experience in creating trust for digital services has taught some important lessons, such as the need to adjust to different individual beliefs and emphasise the need for understandable and straightforward services where data privacy and data usage are explained. Moreover, it is enforced that personal data is used only for providing a better service.

Empowering: In exchange for trust, the service offers an added value that traditional one-dimensional services have been unable to. The user is viewed as an active creator, not a mere passive receiver of a government service, allowing higher effectivity (instant real-time service provision), aversion of risk behaviour (predictive health data allowing you to prevent future illness thanks to early discovery), or new employment).

Making the future happen: Riisalo says this new era of government is coming based on making the most of the information that government holds. He highlighted the case study of the Estonian state portal, which provides a comprehensive resource for Estonian citizens and residents. Citizens don’t need to figure out which government authority provides which service, but instead services are grouped as they are really used by people, under topics like family and business.

Watch this space: Riisalo says Estonia has the authorisations in place to drive this reform, and GGF will be keeping an eye on implementation in the months ahead.

Upcoming Global Government Forum events

  • GGF webinar today: Unlocking insight from data to support decision making
  • GovernmentDX: How the US federal government can successfully transform services in the digital age will be the focus of a new Global Government Forum conference held on 18 and 19 April at Ronald Reagan Building and International Trade Center, Washington, D.C. Sessions include ‘Unlocking insight: how government can make the most of analytics and AI’. Find out more and register to attend here.

Trust in digital ID? That depends

Trusted and verified: Findings of a UK report into public trust in digital identity services show that the more information people have about digital ID, such as what it is, how it works, and what it is for, the more people trust it.

Perceptions shift: The report by the UK Department for Science, Innovation & Technology (DSIT) draws on five online workshops held between April and June 2023. It showed that participants’ attitudes towards identification and digital ID shifted as they understood more about it. “As participants’ discussions developed, many began to think of identity documentation as a basic human right,” the report said.

Three takeaways: The three key findings were that trust in a digital ID depends on trust in government, that people need to trust that the data they share is protected and must have assurances that service providers are not solely motivated by money.

Minister weighs in: In a statement on the UK government website, Saqib Bhatti, the UK minister for tech and the digital economy, gave a sense of what was to come following the report. He said that in the coming year, the government would set up processes “to make real the rules and legislation we’re putting in place”, adding that the findings from the report on trust would inform this work “at every stage”.

How government can build digital ID trust: The development of digital identities is one of the key building blocks for the government of the future. Global Government Forum research has identified digital ID systems as an essential capability in realising visions of more seamless services wrapping around the user, and has covered how three governments are developing digital ID

Busting myths: However, governments have also noted that there is some vocal pushback against these plans, with the UK government having previously been moved to publish a myth-busing article over its plans to develop digital identify verification for online services.

Global Government Forum will hold a webinar on 20 June looking at how government can tackle the objections people have to digital ID, and how it can prove the usefulness of digital IDs in providing more integrated services. Register here.

Thanks for reading: This is the first Global Government Forum AI Monitor. Please get in touch with the team if you have stories to share.

Did you like this roundup? Sign up for Global Government Forum’s email to receive the next Government AI Monitor.

About Jack Aldane and Richard Johnstone

Leave a Reply

Your email address will not be published. Required fields are marked *