Back to search results

Joint paper also underlines need for global cooperation

Artificial intelligence needs ‘the right balance of trust’ between human operators at nuclear plants and AI systems.

The successful integration of artificial intelligence (AI) into nuclear operations will depend on striking “the right balance of trust” between human operators and AI systems, nuclear regulators from the UK, US, and Canada have said.

In a joint paper on AI in the nuclear sector the regulators underlined the need for international cooperation to harmonise regulatory approaches as AI technologies evolve.

The paper marks the first international collaboration between regulators to address the use of AI in the industry, the UK Office for Nuclear Regulation (ONR) said.

The ONR, the US Nuclear Regulatory Commission (NRC), and the Canadian Nuclear Safety Commission (CNSC) co-authored the paper, which outlines how AI could benefit nuclear safety, security and safeguards.

The paper says AI can be a tool to analyse large datasets with the goal to manage risks, improve efficiency, and reduce worker exposure in hazardous environments.

Among its conclusions, the paper says that the rapid pace of AI development means that “it is unlikely that AI-specific consensus standards for the nuclear domain will be available to support regulatory activities within the near future”.

In the meantime, existing nuclear-specific standards should be adapted to address the unique attributes of AI technologies, the paper said.

According to the paper, in the nuclear industry, the quality and integrity of data will be crucial for AI system performance. Data deficiencies, such as bias or inaccuracies, can lead to faulty AI outputs with serious consequences, it warned.

To ensure successful AI integration, meticulous data curation, validation, and governance throughout the AI lifecycle will be essential.

The paper addresses other topics, including integrating AI into existing nuclear systems, managing the AI lifecycle from design to deployment, and common areas of focus for regulators when approaching AI, including considerations for safety cases for AI in nuclear applications.

Shane Turner, technical director at the ONR, said the regulatory collaboration will support the global nuclear community in understanding critical considerations for AI deployment.

“ONR is open to innovation and committed to enabling the safe and secure use of AI and other innovative technologies in the nuclear sector,” Turner said.

Date: Thursday, 26 September 2024
Original article: nucnet.org/news/nuclear-regulators-call-for-right-balance-of-trust-with-human-operators-9-3-2024