We consider that AI has the potential to deliver many benefits to the nuclear industry, such as increasing efficiency or reducing risks to people and the environment. However, challenges remain in using AI safely and securely, which need to be managed.
We are actively looking to strengthen our regulatory arrangements through working with dutyholders who are considering the deployment of AI on nuclear sites, where the benefits can be clearly articulated. Please contact us for more information.
What we have done
- We have consulted with a wide range of stakeholders, including dutyholders, other regulators, academia and government, internationally and both inside and outside the nuclear industry, informing our position on the challenges of AI and how best to regulate it.
- We have undertaken research to determine whether our current regulatory approach and the underlying legislation are suitable for regulating AI.
- We have used regulatory sandboxing to test our arrangements in two AI applications, to determine where the challenges are likely to be and to encourage further developments in the use of AI.
- We have responded to the government white paper on the regulation of AI to outline our approach to regulating AI.
- We have collaborated with the CNSC and USNRC to develop joint principles for the regulation of AI.
What we are doing
- We are working with the IAEA and other national regulators to produce guidance on the safe use of AI.
- We will actively encourage the deployment of AI on nuclear licenced sites to ensure the benefits of AI are achieved, and further understand its capabilities and limitations.
- We will work with a range of stakeholders to encourage better understanding of ways in which AI can be designed and implemented to enable demonstration of safety and security in a wider range of applications and where the potential consequences of failure are more significant.