Centre for Public Policy

The Centre for Public Policy hosted the Cabinet Office for a Policy School on Data and AI in March. Over the week, experts from across the University shared their insights and research with Cabinet Office colleagues to build an understanding of AI and how it could impact public policy. The Centre provided oversight throughout on the policy implications, considerations and opportunities of AI alongside civil service colleagues. 

The week started by providing participants with a baseline understanding of AI development, risks and opportunities. The overall message from the week was that there is a huge opportunity for AI to be used as a tool by policy professionals, but there is a critical need for accountability, transparency and fairness in its development. 

Setting the scene

Professor Ana Basiri from the Centre for Data Science and AI guided colleagues through the first day of the school with an introduction to the topic. She outlined the role the government has in regulation, risk management, values and communication.

Professor Alessandro Vinciarelli, School of Computing Science, provided an overview and history of AI, from the 1940s right through to Deep Learning and modern AI. He debunked some common-misconceptions and fears about AI development and gave an optimistic view of the future.

Moving into the application of AI, Professor Dame Muffy Calder, Head of College of Science and Engineering, gave a real-life example of how a machine learning model for prediction could operate using a hypothetical example. This provided Cabinet Office colleagues with a tangible understanding of how this technology can be used for prediction purposes.

Mitigating Risks

The school then moved into discussing an important consideration for all policymakers: the potential risks and bias of AI. Dr Simone Stumpf, School of Computing Science, provided an insightful presentation for the policymakers on responsible AI. This was an essential element of the learning about mitigating the risks of this technology especially for public service delivery. She emphasised the importance of human involvement at every step of the AI development pipeline, underlining the importance of accountability, transparency and fairness.

Dr Mark Wong, School of Social and Political Science, provided an important input on tackling racial bias and harms. He discussed with attendees how to build trustworthy AI with minoritised ethnic communities using co-design principles. He gave real life examples of how AI can perpetuate racism through discrimination, stereotypes and misinformation.​

A tool for policymakers

It was becoming clear that AI could be used as a tool by policymakers in their own day-to-day work. Professor Alison Heppenstall, School of Social and Political Studies, spoke about the opportunities in creating synthetic populations for use in policy design. This offers policymakers the opportunity to test the potential long-term implications of policy interventions. Alison highlighted the work of the SIPHER consortium and how it had mapped the impact of child poverty interventions on health.

Professor Marion Scott and Professor Claire Millar both spoke from the School of Mathematics and Statistics. They discussed the power of data as an underused asset for informing policy. They highlighted how AI can be used to harness multiple data sources and facilitate more informed policy decisions. 

Our Director, Professor Nicola McEwen and Professor Ana Basiri rounded off the day by challenging participants to think about the policy implications of everything they had heard on the development, potential and risks of artificial intelligence. 

 

 

Field Visit to Wolfson Wohl Cancer Research Centre

Participants in the Policy School were also given the opportunity to visit the University’s Wolfson Wohl Cancer Research Centre in Bearsden. This was expertly facilitated by the Government Relations Team at the University.

This visit enabled participants to hear from leading academics at the institution working in the field of AI and health. It also provided a chance to hear directly from clinicians working in the NHS about their experience and how these advances could improve cancer diagnoses. They heard about the practical application of AI in healthcare settings, opportunities and the potential for further innovation. This provided a tangible example of how policymakers could be supporting AI development for the delivery of healthcare services going forward.

Cancer research and innovation

Following a tour of the labs, Professor Nigel Jamieson provided insights into how his team use spatial transcriptomics approaches to investigate cancer and inflammatory diseases. He underlined the opportunity AI presents for enabling practitioners to analyse significantly larger datasets in much more minute detail than before. He noted it’s about making big data, useful data.

Professor John Le Quesne walked the attendees through his group’s work on new prognostic and predictive biomarkers in lung cancer. He discussed how these innovations can be used to improve cancer treatment.

Professor Joanne Edwards spoke about the INCISE project she leads, aiming to transform bowel cancer screening by developing a tool that can predict which patients with pre-cancerous growths in their bowels, called polyps, will develop further polyps. She highlighted how these innovations could transform diagnosis and the patient experience.

The end result

Throughout the week there were also insightful presentations from policy colleagues in the civil service on evaluation, the policy process and wider policy context of AI. This coupled with the real-life examples of AI application and research provided participants with specialist expertise on the role of AI and data in policy. Groups ended the week by presenting their policy solutions on AI and data in public services to a panel of expert judges from across the University and civil service, this included our Director, Professor Nicola McEwen. They harnessed their learning from across the week to set out the opportunities, risks and potential areas for government investment and intervention.

The team at the Centre for Public Policy were delighted to collaborate with the Cabinet Office on this Policy School. Kimberley Somerside, our Communications and Policy Engagement Lead, worked closely with Paul Bradley, Head of Government Relations, to design and facilitate the inputs from across the University during the week. It provided an opportunity to share learning between academia and policymakers about a hugely important and topical issue for governments. We hope to build more connections between the research and policy community through projects such as this and are grateful to everyone who contributed to the week.

 

 

Policy Engagement Opportunities on AI

The Westminster Public Accounts Committee has recently launched an inquiry on the use of AI in Government which is accepting written evidence until Friday 3 May.

The National Audit Office (NAO) reported in March 2024 that 70% of government bodies surveyed are piloting or planning the use of AI, including for supporting operational decision-making and improving internal processes. The call for evidence covers:

• Departmental accountability on AI delivery, funding and implementation;
• Progress on strategy development and governance arrangements;
• Risks and opportunities of AI adoption in government;
• Data and skills issues in government.

More information on the inquiry can be found here.

  

---

For more information about the Centre’s policy engagement work, please contact Kimberley Somerside, Communications and Policy Engagement Lead at the Centre for Public Policy.  

First published: 10 April 2024