MaThRad Annual Review & Community Mixer 2025
MaThRad Annual Review & Winter Research Mixer 2025
Warwick's Nuclear Mathematics & AI Community of Practice
15th December 2025 | 10:30 am - 2:30 pm | MSB 2.22, Mathematical Sciences, University of Warwick
The event brings Warwick’s Nuclear Mathematics & AI Community of Practice together and welcomes colleagues from across campus to engage with MaThRad’s work in the mathematical theory of radiation transport across fission, fusion, and nuclear medicine. It offers space to connect across these areas, share current work, and think together about the year ahead.
Programme (10:30 – 14:30)
| Time | Session |
|---|---|
| 10:30 – 12:00 | MaThRad Year-in-Review [Closed Group - MaThRad Researchers Only] Overview of key scientific highlights, outputs, collaborations and milestones from the past year. |
| 12:00 – 12:45 | Panel Discussion [Open to All] Governing Agentic AI in the Nuclear Age: Where Mathematics, Experience, and Power Meet A chaired discussion with Prof. Paul Smith, WMG and MaThRad colleagues exploring mathematical, scientific and governance challenges in nuclear modelling and AI. |
| 12:45 – 14:30 | Networking Lunch & Festive Refreshments [Open to All] |
Panel Discussion
Governing Agentic AI in the Nuclear Age:
Where Mathematics, Experience, and Power Meet
Agentic AI is no longer something the nuclear sector can treat as emerging or experimental. It is already entering workflows, shaping procurement choices, and being built into the infrastructure that will support the next phase of nuclear and AI development. At the same time, private capital and hyperscaler interests are actively securing the compute, data and energy systems needed to keep this acceleration going.
The risk is not simply that AI arrives, but that it arrives at pace, packaged as efficiency or augmentation, before there has been time to ask what forms of judgement, responsibility and control we are prepared to retain and what we may be giving up by default.
This session is deliberately exploratory. It is not about presenting a finished governance framework or offering solutions in advance. Instead, it creates space to surface the questions that are easiest to skip when technologies move quickly and incentives are already aligned.
Bringing together perspectives from industry, nuclear modelling, data assimilation, probability and discrepancy, cyber security, and critical studies of AI, with relevance across other safety-critical domains, the discussion will work through questions such as:
-
Where should agentic AI never be given the initiative in nuclear systems — and who has the authority to draw that boundary?
-
How can tacit expertise, safety culture and technical agency be protected when AI enters under the banner of optimisation or workforce augmentation?
-
What kinds of mathematical, organisational and institutional evidence would regulators, engineers and operators actually need before trusting agentic AI in safety critical settings?
The aim is not consensus. It is to leave with sharper questions, visible tensions and clearer lines of enquiry, so that Warwick’s nuclear mathematics and AI community is better prepared to engage with what is already unfolding.
Contributors
Professor Paul Smith
Visiting Professor at Imperial College London and Bangor University, with over four decades of experience as a mathematical modeller in the nuclear industry.
Dr Anita Khadka
Assistant Professor in Trustworthy and Responsible AI at the Warwick Manufacturing Group (WMG)
Professor Andreas Kyprianou
Professor of Probability in the Department of Statistics at the University of Warwick.
Professor Sanjay Sharma
Sanjay Sharma is Professor at the Centre for Interdisciplinary Methodologies, University of Warwick.