How We Are Using AI

image of a laptop with icons overlaying within a person typing on keyboard

On this page:

Responsible and Ethical Use of AI

To help achieve the agency's important safety and security mission, the NRC is responsibly incorporating advanced technology—such as artificial intelligence (AI)—into the agency’s daily work. This allows NRC staff to access and analyze information more efficiently, make better-informed decisions, and communicate more clearly throughout the organization as well as with the public and its stakeholders.

When using AI, the NRC follows strict ethical standards. AI tools are used to support, not replace, the judgment and expertise of NRC staff. Decisions about nuclear safety and security remain in the hands of experienced professionals who use AI as one of many resources. The NRC reviews AI-generated outputs for factual accuracy, bias, and explicability. If there are uncertainties or gaps in AI-generated information, these are clearly communicated, and further review is conducted when needed.

The NRC is committed to continually improving its use of advanced technology. By responsibly integrating AI, the NRC strengthens its ability to meet its mission, uphold its principles and core values, and comply with legislative requirements for the accelerated use of efficiency-enhancing AI capabilities.

To top of page

Leveraging AI for Efficiency

Our approach to AI centers on making routine tasks, such as researching official records and creating new content, more efficient, thereby freeing up staff time for more complex work. To this end, we have established AI capabilities that are tuned to the mission and functions of the NRC and that reference curated documents and institutional knowledge.

Using internally-developed tools, documents and search models are organized in a fashion that promotes collaboration and knowledge sharing. Input documents are carefully curated by subject matter experts across the agency who ensure the quality, relevance, and timeliness of the reference material. This helps to ensure that the outputs of AI tools are grounded in trusted information.

In support of well-defined, routine processes, predefined prompts and templates guide interactions with AI, helping ensure consistency and accountability in how information is retrieved and referenced while also controlling the form and format of AI-generated outputs. Staff are empowered to use these tools to enhance productivity and to explore innovative applications. This approach reflects our ongoing commitment to responsible innovation and the ethical adoption of new technologies in support of our regulatory mission.

To top of page

AI Use Case Inventory

Consistent with Executive Order (EO) 14179, the NRC continues to update its inventory of non-classified, non-sensitive, and non-research use cases of AI. The following use case(s) reflect the agency's approach to AI as described above.

SimplifAI:  An advanced AI assistant developed by the U.S. Nuclear Regulatory Commission (NRC) to deliver clear, logically organized, and in-depth analysis of authoritative regulations, guidance, regulatory inputs, and relevant scientific and technical sources across a wide range of NRC programs and regulatory functions. SimplifAI streamlines information retrieval and synthesis, delivering timely responses to technical, regulatory, and procedural queries. This enables NRC staff to focus their expertise where it is most needed, supports efficient project management, and reduces the time spent on routine (though often extensive) research, thereby maximizing the agency’s resource effectiveness. SimplifAI is explicitly aligned with the NRC mission as well as its five Principles of Good Regulation—Independence, Openness, Efficiency, Clarity, and Reliability. The system is not used as a substitute for professional judgment or legal interpretation but acts as an enabling tool that enhances the NRC staff’s ability to carry out their regulatory responsibilities with greater confidence, efficiency, and transparency.

To top of page

Page Last Reviewed/Updated Tuesday, September 9, 2025