A team from the University’s Department of Mathematics and Institute for Mathematical Innovation were in Parliament last week as part of ‘Evidence Week’, organised by the charity Sense about Science.

Exploring ‘Black Boxes of AI’ – through the mathematics that underpins algorithm powering AI technologies – the team met a range of MPs, Peers and advisors to brief them on their work, activities and the role of maths in helping us all navigate increased digitisation.

Mathematics for Deep Learning (M4DL) combines theory, modelling and computation to unlock the next generation of deep learning with a transformative potential across all aspects of our life. It is a collaboration between Bath, Cambridge and UCL, funded by EPSRC.

The Institute for Mathematical Innovation (IMI) at Bath adds value and impact to research by creating interdisciplinary connections through modelling, data science and machine learning. It focuses on themes, including the application of maths in response to environmental extremes, health and quantum technologies.

MPs who attended the session in Parliament’s Jubilee Rooms included Chi Onwurah, Chair of the Science, Innovation & Technology Committee, and local MP for Thornbury & Yate, Claire Young – herself a mathematics graduate.

Peers included former Defence Secretary Lord Browne of Ladyton, Lib Dem Lords Spokesperson for Science, Innovation & Technology Lord Clement-Jones, as well as leading science communicator and broadcaster Baroness Freeman of Steventon.

The team discussed AI-related questions parliamentarians are facing, helping to shed light on some of the systems which power new technologies such as ChatGPT. Conversations also considered opportunities through AI, as presented in the recent AI Opportunities Action Plan.

In addition, they presented a policy brief summarising key aspects of their work to date. This makes calls to ensure mathematically based tools are embedded to minimise unwanted bias in the training of future AI systems, and a focus on the importance of mathematics skills at all levels.

Prof Chris Budd OBE who led the team said: “It was an honour to take such a great team to Parliament to discuss the very important issues related to AI with so many MPs and Peers. It is clear that this topic is of great interest and concern to those in Government and Parliamentarians as AI is going to change our lives in many ways. We hope that the insights provided by mathematics will help to make AI more transparent and will reduce some of anxiety associated with its numerous applications.”

Dáire O'Kane, PhD Research Programme in Mathematics at the University of Bath, added: “Our trip to Westminster was a fun and exciting opportunity to raise awareness to MPs and Peers about the important role of maths in not only understanding AI but also in explaining its limitations. I particularly enjoyed discussing work on deep fake detection and the ‘magic’ behind generative AI, as well as promoting the adoption of differential privacy techniques when working with sensitive data.”

Tracey Brown OBE, Director of Sense about Science, said: “Parliament has the job of scrutinising policies, legislation and emerging issues across a mind-boggling array of subjects – few perhaps more pressing than the growing use of AI. The Institute for Mathematical Innovation and team from M4DL provided a crucial component of Evidence Week in Parliament. The team came prepared to brief MPs, peers and parliamentary staff at all levels of understanding and the constant buzz around their briefing pod showed how successful they were.”

The full team in Parliament included: Chris Budd OBE, Principal Investigator, Maths4DL; Yolanne Lee, PhD candidate, UCL; James Rowbottom, PhD candidate, University of Cambridge; Cangxiong Chen, Research Associate, University of Bath; Dáire O’Kane, PhD candidate, University of Bath and Jenny Power, PhD candidate, University of Bath.

The team were also joined in Parliament by Andy Dunne, University of Bath Head of Public Affairs.