AI4Nuclear would like to invite you to a series of online workshops to define the requirements for a national strategy for Artificial Intelligence. The proposal for this strategy is set out in the attached document which will form the basis for discussion at this workshop. The outputs of the workshop(s) will be a policy driven paper to inform UK governments future strategy around artificial intelligence and digital to modernise approaches to whole life safety and regulation meeting the recommendations of the Fingleton led nuclear regulatory review, published in 2025. We will be concluding this activity at the AI4Nuclear conference at STFC Daresbury in October 2026.
The focus will be on the following areas:
(1) People, skills and culture:
· Skills development focusing on enhancing existing competencies in nuclear engineering and sciences. Define a sector-wide AI competency framework (e.g., AI model developers, AI assurance engineers, human–AI teaming specialists).
· New skilling of individuals working in the sector through for example the creation of a nuclear sector focused AI Skills Boost programme. Making sure that we have the right skills to right problem building on the current technocratic nature of the sector. Establish professional accreditation routes via to formalise AI capability in nuclear.
· Introduce leadership training on AI governance and strategic adoption to drive culture change and accountable ownership across the sector.
(2) Governance, ethics and psychological safety:
• Ensuring organisations have the correct policies, procedures and guidance to implement AI within UK and international regulatory and legal frameworks.
o Define procedures for application of AI in specific business contexts covering topics including managing model drift, dataset degradation, bias detection/mitigation, and retirement criteria.
• Ensuring that individuals have the confidence to report issues accurately in an industry where AI enhanced decision making (and automated) decision making becomes increasingly common. This is vital to ensuring that AI (particularly agentic AI) has trustworthy data from which to base models on.
o Codify human-in-the-loop / human-on-the-loop principles for safety-relevant and regulator-facing AI applications.
o Define psychological safety metrics and reporting channels, especially where AI outputs conflict with human judgement.
(3) Data and knowledge management:
• Ensure that UK sovereign data can be exploited by UK companies to maximise the GVA generated from our 80+ years of nuclear knowledge.
• Nuclear-sector data is uniquely positioned to advance AI in areas such as ultra-reliable anomaly detection, robust uncertainty quantification, physics-informed neural networks, safety verification, and explainable systems for high-risk environments. Because these applications demand provable reliability under uncertainty, nuclear operations provide an exceptionally rigorous testbed for developing next-generation AI validation, safety, and assurance frameworks for mission-critical systems.
The format of the workshop will be a Q&A and round table discussion on the topics above where we will ask key questions around:
• What are we good at today?
• What do we need to be good at tomorrow?
• What are the enablers and blockers?
Click here to see co-ordinating document.
When: Wednesday 24th June, 10:00-12:30 (UK time)
Where: MS Teams (access will be provided in the registration confirmation email you will receive.)
Registration Deadline: Wednesday 24th June, 10:00