ML25273A290
| ML25273A290 | |
| Person / Time | |
|---|---|
| Site: | Nuclear Energy Institute |
| Issue date: | 09/24/2025 |
| From: | Slider J, Voelsing K Nuclear Energy Institute |
| To: | Office of Nuclear Regulatory Research |
| References | |
| Download: ML25273A290 (1) | |
Text
©2025 Nuclear Energy Institute NEI Perspective NRC AI Workshop September 24, 2025 Jim Slider jes@nei.org Kelli Voelsing khv@nei.org Nuclear Industry Use of AI
©2025 Nuclear Energy Institute 2 Topics CURRENT NEI ACTIVITIES
- 1. Facilitating Engagement with NRC on AI issues
- 2. NILLM - Nuclear Industry Large Language Model
- 3. NEI Regulatory Requirements Checking Tool Piloting an AI-enabled tool for NEI members Enabling efficient access to and application of industry knowledge and data
©2025 Nuclear Energy Institute 3 Adoption Rapid and accelerating adoption Strong interest and hundreds of use cases across all plant departments Many uses cases focus on
- Efficient access to the right information to improve products and support informed decision making
- Automation of administrative or routine processes Human-in-the-loop (HITL) is used for validation and decision making when utilizing AI Regulation Large value to be gained from uses that do not require NRC approval Nearer-term regulatory submittals likely to involve
- Image recognition and processing
- Machine learning and pattern recognition Regulatory applications involving Generative AI (Gen AI) are likely a bit further out A graded and flexible regulatory framework is critical
-Many kinds of AI
-Large spectrum of how AI is used in any application
-Encourage innovation that benefits safety and efficiency
- 1. Facilitating Engagement with NRC on AI issues
©2025 Nuclear Energy Institute 4 AI-READY NRC DATASETS
Partnering with NRC staff to deliver high-value, AI-ready NRC datasets
Promotes efficiency and transparency
Supports all stakeholders
Feedback to ensure
Sustainable process
Right metadata and formats
Priorities for future datasets
Initial datasets include
Regulatory Guides
- 1. Facilitating Engagement with NRC on AI issues
PURs
LR/SLRs NRCs collaboration and support on this effort is extremely valuable
IMCs/IPs
SECYs, etc. etc.
©2025 Nuclear Energy Institute 5
- 2. NILLM - Nuclear Industry Large Language Model A Collaborative Industry Project Initial use case testing begins Q4 2025 Project Sponsor - Ramon Martinez, INPO (martinezr@inpo.org)
©2025 Nuclear Energy Institute 6 Initial Use Cases & Scope Knowledge Transfer (led by NEI)
Supports knowledge management and transfer by enabling overview and detailed understanding of key topics in the nuclear industry Operational Experience Search (led by INPO)
Smart search across IRIS, INPO, EPRI, and NEI documents Supports plant reliability and safety initiatives Troubleshooting (led by EPRI)
AI-assisted plans for degraded components or system conditions
©2025 Nuclear Energy Institute 7
AI-enabled Document Compliance Check
Automated identification and curation of regulatory requirements for a given application
Assess how well a document addresses each requirement
Supports high-quality and consistency in regulatory submittals
Useful for self-assessment of regulatory compliance
Near-term, 6-month pilot, proof-of-concept project
What will we learn from the pilot?
Increase comfort with AI-enabled regulatory and licensing applications
Understand appetite and adoption of targeted AI-workflows versus open-ended chatbots and prompt engineering
Assess value to end-users
Hosting and usage costs
- 3. NEI Member Pilot AI-ENABLED REGULATORY REQUIREMENTS TOOL Alignment Focus Insight NEI is partnering with Jensen Hughes to make this tool available to NEI members Demonstration of the tool later on todays agenda
©2025 Nuclear Energy Institute 8 Closing Thoughts Many current uses of AI in the nuclear industry are outside NRC requirements:
- Efficient access to information
- Automating administrative tasks NEI supports development of AI regulatory framework that is:
- Flexible
- Graded NEI and the nuclear industry are committed to:
- Practical use of AI for the benefit of safety and efficiency
- AI that is ethical, responsible, explainable, verifiable