The Office for Nuclear Regulation (ONR) is set to explore the integration of artificial intelligence (AI) into the UK nuclear sector through a new regulatory sandbox, supported by £3.6 million in funding from the government’s Regulatory Innovation Office (RIO). This initiative, which aims to enhance safety and efficiency, will run until March 2026 and marks a significant step for the independent regulator responsible for overseeing nuclear safety and security in the UK.
Paolo Picca, Innovation Lead at ONR, emphasized that the funding was secured as part of the RIO’s AI Capability Fund, which was established to streamline regulation in key growth industries. “Discussions with the nuclear industry highlighted a pressing need to evaluate the opportunities that AI can bring,” he stated. The competitive bidding process allowed ONR to partner with other regulatory bodies, including the Environment Agency and the Defence Nuclear Safety Regulator, thereby fostering a united approach to AI deployment in the nuclear sector.
With the UK government keen on leveraging technological advancements for economic growth, ONR’s project seeks to promote collaborative discussions across various sectors of the nuclear industry. Picca noted, “The overall goal is to ensure the safe deployment of AI in nuclear installations, thereby supporting the government’s ambitions for efficiency and growth.” As part of this effort, ONR aims to avoid disparate regulatory approaches and establish a unified framework for AI implementation.
The potential applications of AI in nuclear regulation are vast. Picca described the current regulatory landscape as “technology-neutral,” allowing for innovations that can improve safety and efficiency. The sandbox initiative will explore machine learning (ML) applications, particularly in areas like computer vision and data management. “We are currently assessing the use of ML for classifying operational data and utilizing large language models (LLMs) for internal knowledge management,” he explained. Notably, the ongoing trials include efforts to simplify regulatory guidance documents and support knowledge transfer among inspectors.
Two key test cases have been identified for the regulatory sandbox: nuclear waste characterization and non-destructive testing. Picca highlighted that feedback from the industry led to the selection of these areas for their potential to enhance efficiency and safety. “AI can significantly improve the classification of waste, potentially reducing long-term storage costs,” he said. Similarly, the use of AI in non-destructive testing could streamline inspection processes, reducing both time requirements and occupational exposure for workers.
As the sandbox initiative progresses, ONR is also taking lessons from its previous experience with a pioneering regulatory sandbox pilot conducted in 2022-2023. Picca noted that this pilot generated interest from global regulatory agencies and helped inform the current project. “The regulatory sandbox allows for collaborative exploration of new technology ahead of formal assessments,” he stated, underscoring its role in reducing uncertainty around innovative solutions.
Following the confirmation of funding, ONR has engaged stakeholders through workshops and events to establish a foundational understanding of AI design and assurance. Progress is being made on both test cases, with workshops scheduled to deepen discussions on the challenges and opportunities at hand. “While we have encountered no major technical issues thus far, maintaining our timeline is crucial as we aim to publish a summary report in spring 2026,” Picca said.
Despite the promise of AI, challenges remain regarding its broader implementation in nuclear regulation. Picca noted the industry’s responsible approach towards AI, emphasizing collaboration with international peers to develop sector-specific guidance and standards. Ongoing discussions with nuclear regulators in the US and Canada have informed a position paper outlining international principles for regulating AI in the nuclear sector.
In terms of fostering an innovative mindset within the nuclear industry, Picca highlighted recent dialogues among innovation leaders, which identified mindset as a barrier to progress. He emphasized ONR’s commitment to encouraging open discussions and clarifying regulatory challenges to facilitate innovation. “Our goal is to clarify why particular solutions may not be applicable, which can ultimately help the industry innovate further,” he noted.
As the regulatory sandbox nears its conclusion in March 2026, ONR envisions a future where UK nuclear licensees increasingly engage with innovative solutions. The establishment of a handbook for future sandboxing activities is part of ONR’s broader aspiration to reduce regulatory uncertainty surrounding new technologies. Picca concluded, “ONR is open to innovation and committed to enabling the safe deployment of AI and other technologies in the nuclear sector, fostering an environment where innovation can thrive.”
See also
CISA Releases AI Integration Guide for Operational Tech, Emphasizes Safety and Security
AI’s Growing Energy Demand: Companies Must Elevate Governance to Mitigate Risks
DeSantis Opposes Trump’s Federal AI Regulation Push, Advocates for State Control
India’s AI Governance Guidelines Transform Auto Sector, Shifting Liability to Manufacturers
Europe’s AI Ethics Platforms Market Forecast to Soar to $45.3 Billion by 2035 Amid Regulatory Shift
