China is integrating artificial intelligence into its nuclear energy sector with the ADANES system, aiming to enhance safety and efficiency throughout the reactor lifecycle. However, concerns remain about the transparency of AI 'black boxes' in such critical applications.
China's ambitious integration of AI into nuclear power raises critical questions about the balance between technological advancement and safety, particularly given the opacity of current AI systems and the immense risks associated with nuclear energy.
China is advancing its integration of artificial intelligence into the nuclear energy sector with the unveiling of the Accelerator-Driven Advanced Nuclear Energy System (ADANES) by researchers at the Chinese Academy of Sciences (CAS). Presented at the World Artificial Intelligence Conference in Shanghai, ADANES is designed to permeate the entire nuclear energy lifecycle, from design and commissioning to operation and maintenance, with the goal of fundamentally altering safety logic and enhancing quality, efficiency, and safety.
Proponents, like Wang Shoujun, president of the Chinese Nuclear Society, view AI integration into the economy, including nuclear power, as inevitable. They argue that proactive planning and safety measures are crucial. The ADANES system's architecture comprises five layers: unified data infrastructure, physics-native world models, physical-system control, intelligent-agent coordination, and continuous evolution. China is also developing a national platform for engineering verification to ensure the system's stability.
However, the opacity of current large language models, often referred to as 'black boxes,' presents a significant challenge to the stringent transparency requirements of nuclear energy safety. This lack of understanding is at odds with the need for absolute certainty in managing nuclear reactors. Meanwhile, in the United States, Microsoft and NVIDIA are collaborating on AI-powered toolkits aimed at streamlining the lengthy and costly processes of permitting, design, and engineering for new nuclear plants, utilizing generative AI for tasks like document drafting and gap analysis.
The broader AI boom is also a catalyst for developing new nuclear energy technologies, with significant investment from Silicon Valley aimed at powering the substantial energy demands of generative AI. This push for advanced nuclear capacity is occurring globally, with some U.S. startups also reportedly entering the sector with less emphasis on safety measures.