The Office for Nuclear Regulation (ONR) and the Environment Agency have held their final stakeholder panel session to examine the end-of-project findings from the artificial intelligence (AI) regulatory sandboxing pilot.
Up to 60 people attended the collaborative event to share initial conclusions from this pioneering trial scheme, understood to be the world’s first ever regulatory sandboxing within the nuclear industry.
The regulation of AI has been scrutinised through the lens of two example applications – the operation of a robotically operated glovebox and its use in supporting and challenging structural integrity claims.
A report will now be compiled summarising the findings from the pilot which will be presented to the government’s Regulators’ Pioneer Fund (RPF), who provided a grant to help finance the project for both ONR and the Environment Agency.
The RPF enables UK regulators and local authorities to help create a regulatory environment that encourages business innovation and investment. The current £12m round is being delivered by the Department for Science, Innovation and Technology.
At the recent meeting, attendees included the National Nuclear Laboratory, academia, other UK regulators, nuclear industry licensees, the Nuclear Institute and representatives from government departments.
ONR will continue its sandboxing work beyond the end of this pilot and is keen to progress this approach in other areas of innovation where it is in the interest of society and consistent with safety, security and safeguards expectations.
David Smeatham, ONR’s Head of Innovation, said: “I’m grateful for all the support that stakeholders have given to this work.
“It wouldn’t have been possible to paint such a rich picture of the opportunities and factors associated with the use of AI in nuclear without everybody’s valuable expertise and inputs.
“It was an extremely beneficial and enjoyable piece of collaborative work – thank you to all those who engaged with us.”
Jake Surman, Senior Advisor RSR Research, Development & Innovation Lead, from the Environment Agency, said: “Our joint regulatory sandbox pilot has been an excellent opportunity to develop a single, joined-up regulatory engagement for nuclear innovations of mutual interest.
"The sandbox pilot created a collaborative space between regulators, industry and the supply chain to take considerations around the nuclear sector use of artificial intelligence out of the abstract and explore, in practice, regulation of this technology.
"Not only will the output from the pilot inform our approach to regulating innovation, it will also contribute to our future regulation of artificial intelligence in the sector.”
Timothy Burd, Senior Consultant, from Adelard, part of NCC Group, said: “We are delighted to be supporting this work by developing mock assurance cases for two case study systems.
“The sandboxing process has created a unique environment where a wide-range of stakeholders can collaborate and problem solve together, which has enabled us to make significant progress in developing an approach for this important area.”
Findings from this project will be reflected in other ongoing stakeholder work including studies being undertaken by the Nuclear Institute’s AI4N initiative, the development of high level regulatory principles with other international nuclear regulators and the development of the concept of an international regulatory sandbox, with the benefits that increased harmonised approaches could provide.
The UK pilot scheme, made possible by a grant from the RPF, awarded by the Department for Business, Energy & Industrial Strategy (BEIS), explores how the use of AI could be regulated in a way that realises its potential benefits within the nuclear industry and enables innovation where it is in the interest of society and consistent with safety, security and environmental protection expectations.
The work also benefitted from funding provided by the Control and Instrumentation Nuclear Innovation Forum.
Sandboxing offers regulators and industry a safe space to consider potential challenges and solutions, or where more work is required.