The Diablo Canyon Power Plant, California’s only remaining nuclear facility, is set to implement a generative AI platform developed by Atomic Canyon to streamline document retrieval. Announced by Atomic Canyon and Pacific Gas and Electric (PG&E), the Neutron Enterprise AI will focus on document search and retrieval, not operational functions, leveraging Nvidia AI infrastructure to handle vast amounts of regulatory and technical documentation. PG&E spokesperson Suzanne Hosn noted that efficient access to documentation is “not just a matter of convenience – it’s an imperative” for ensuring regulatory compliance and operational safety.
This AI initiative addresses a persistent challenge in the nuclear sector: scattered and complex documentation. With groups such as engineering, operations, and regulatory compliance relying on accurate records, locating specific documents can take weeks. By integrating Neutron Enterprise’s Retrieval-Augmented Generation (RAG) capabilities, Diablo Canyon aims to expedite this process, reducing the time spent on compliance and maintenance documentation retrieval and potentially setting a new efficiency benchmark in the nuclear industry.
However, generative AI’s broader deployment in nuclear plants remains constrained. The International Atomic Energy Agency (IAEA) has expressed caution, citing the unpredictability and lack of transparency in generative AI decision-making as significant obstacles. Despite these challenges, Atomic Canyon emphasizes that Neutron Enterprise’s use of RAG architecture ensures responses are grounded in specific, verifiable nuclear industry documents, minimizing the risk of misinformation or “hallucinations” common in other AI systems.
This initial deployment at Diablo Canyon could pave the way for broader applications of AI in the nuclear industry, though Atomic Canyon is taking a cautious approach. The company has indicated plans to expand AI applications for document management in design, engineering, and maintenance phases, depending on the success of this project. With Diablo Canyon’s future beyond 2030 uncertain, the plant’s ability to implement AI-driven efficiencies may shape how long it remains operational in California’s shift toward renewable energy.
