GenAI on the rise: UK public sector’s AI appetite and attitudes revealed in survey
More than half of UK public servants in finance, transformation and digital jobs are using generative AI (‘GenAI’) in their day-to-day work – but more than 40% have encountered errors in the course of their work caused by the use of GenAI tools – according to a survey by Global Government Fintech.
Generative AI refers to the use of AI to create new content such as text, audio, images and video that can appear like it has been made by a human. Examples include ChatGPT, a chatbot developed by US-based company OpenAI that uses natural-language processing to understand and generate human-like text in response to user prompts; Copilot, an AI service integrated into Microsoft applications; and Google’s AI tool Gemini.
The UK Public Sector Fintech and Artificial Intelligence Awareness Study was undertaken between 16 July 2025 and 29 August 2025 to better understand public servants’ latest thinking. The online survey, which received 287 responses, asked a series of questions of UK public servants working across central and local government, health and higher education, in finance, transformation and digital roles. The first story on this survey was published last week.
Read more: Anti-fraud, digital ID and payments: UK public sector’s fintech priorities revealed
The AI-focused questions sought to explore the extent to which respondents are using AI tools and their confidence in doing so, as well as views on the wider application of AI in their organisation. The technology is increasingly being championed by most governments worldwide, including the UK where AI is the latest ‘One Big Thing’ campaign focus.
Overall, respondents were positive about GenAI’s use, with more than two thirds believing that the technology has the ‘potential to improve policy development and analysis’ and more than one in two actually using GenAI. But 41.9% have ‘encountered errors in the course of their work caused by the use of GenAI tools’, and 83.6% think they ‘should be actively involved in how AI is introduced to their departments’.
Register now for Public Service Data.AI 18 September 2025 – London, UK
Brought to you by Global Government Forum and hosted by HM Government, Public Service Data.AI is the UK’s flagship annual event for civil servants working to unlock the power of data and artificial intelligence across government. Free to attend for all UK public servants, this event brings together digital leaders, policymakers, data specialists and service designers to explore how the effective use of data and AI can drive smarter, fairer, and more responsive public services.
From modernising digital infrastructure and fostering public trust to ensuring ethical AI procurement and improving data-sharing across departments, Public Service Data.AI 2025 will focus on the key enablers of successful digital transformation. Find out more and register here
Assessing AI confidence
Overall, 50.9% of respondents said they agree and 18.7% said they strongly agree with the statement: ‘Generative AI tools have the potential to improve policy development and analysis’. Just 5.7% said they strongly disagree and 9% said disagree, while 6.7% did not know.
More than 50% of respondents in total strongly agree (19.1%) or agree (37.1%) that they are ‘using GenAI tools such as ChatGPT, Copilot or Gemini in their day-to-day work’. 12.4% strongly disagreed that they were and 10.6% disagreed.
But the numbers of people ‘feeling confident using GenAI in their day-to-day work’ is slightly lower. Just 13.1% strongly agree (feel confident) and 33.9% agree. 20.5% neither agree nor disagree. 15.5% disagree and 8.1% strongly disagreed.
Just under half of respondents in total have actually ‘encountered errors in the course of their work caused by the use of GenAI tools’, with 13% strongly agreeing with this statement and 28.9% agreeing.
The UK government published an ‘AI Playbook’ in February with the aim of giving departments and public sector organisations ‘accessible technical guidance on the safe and effective use of AI’. Created by more than 50 experts from the Government Digital Service (GDS) and the wider Department for Science, Innovation and Technology (DSIT), the playbook consists of input from more than 20 government departments and public-sector organisations, along with insights and peer-review from industry and academic advisers.
But the playbook has yet to have achieved cut-through, it would seem. Just under a third of respondents say their organisation ‘is providing enough guidance or policy on the safe use of GenAI’, with just 6.7% strongly agreeing with this statement and 25.2% agreeing. About the same total either disagreed (18.2%) and strongly disagreed (12.3%).
Read more: GenAI and DLT to have “transformative” impact: Ireland’s digitalisation minister
Wider application of AI
Views on the wider application of AI in organisations were also explored, with respondents asked the extent to which they agreed or disagreed with four specific statements.
Asked first whether they agreed or disagreed with the statement: ‘I understand how AI-based decision systems (e.g. risk scoring, eligibility engines) work in principle’, the balance was – just – in favour of agreement. Over a quarter 27.2% agree and 7% strongly agree, with 21.3% disagreeing and 9.1% strongly disagreeing. 23% neither agreed nor disagreed and 10.5% did not know.
On a second statement – ‘My organisation has a clear process for evaluating and procuring AI-based technologies’ – the balance was slightly negative, with 16.4% disagreeing and 9.4% strongly disagreeing. This was against just 2.8% strongly agreeing and 17.4% agreeing. Overall, 27.2% neither agreed nor disagreed and 24.7% did not know.
Asked whether ‘the senior leadership of my organisation demonstrate sufficient understanding of AI and its implications’, the balance was again slightly negative, with 18.5% disagreeing and 15% strongly disagreeing. 17.4% agreed and just 2.8% strongly agreed. Neither agree nor disagree was the most popular response (28.2%).
The fourth and final question asked the extent to which public servants should be actively involved in shaping how AI is introduced in their departments. This received significantly more positivity, with 36.6% strongly agreeing and 47% per cent agreeing. Just 1% disagreed and 0.7% disagreed.
For more information about our research programme, or to become a research partner, please contact [email protected]
Public sector training opportunities
Global Government Forum provides a range of training seminars on AI in the public sector, including our popular ‘Deploying AI in the Civil Service’, which we can deliver as a bespoke in-house session, tailored specifically for your organisation.
Coming up soon on our open programme we have Mastering Prompt Engineering on Thursday 23 October, and How Artificial Intelligence Can Empower the Civil Service on Thursday 4 December. Details of how to book your place can be found on each course’s webpage.
You can watch back webinars on the subject of Mastering Prompt Engineering (3 September) and Making AI Transformation Happen in Government (9 September).
About Ian Hall
Ian is Global Government Finance editor. He has formerly held roles including UK director of Euractiv (2011-2018), editor of Public Affairs News (2007-2011) and news editor of PR Week (2000-2007). He was shortlisted for ‘Editor of the Year’ at the British Society of Magazine Editors (BSME) Awards in 2010. He began his career in the Balkans at English-language weekly the Sofia Echo (Bulgaria: 1998-1999).
Related Posts
