ML24263A259
| ML24263A259 | |
| Person / Time | |
|---|---|
| Site: | Nuclear Energy Institute |
| Issue date: | 09/17/2024 |
| From: | Slider J Nuclear Energy Institute |
| To: | Office of Nuclear Regulatory Research |
| References | |
| Download: ML24263A259 (1) | |
Text
©2024 Nuclear Energy Institute 1 James Slider Technical Advisor, Innovation &
Regulatory Affairs September 17, 2024 NEI View on Artificial Intelligence
©2024 Nuclear Energy Institute 2
©2024 Nuclear Energy Institute 2 Artificial Intelligence AI is not a new technology AI is growing more capable very rapidly Our culture has not yet adapted, but will The current AI mania feeds two opposite views on AI for nuclear applications:
AI will reduce drudgery and free personnel to focus on matters most important to safety (Robbie the Robot scenario)
AI will threaten plant or public safety (Terminator scenario)
The nuclear industry knows how to adopt new technologies and is doing so with AI Robbie Terminator
©2024 Nuclear Energy Institute 3
©2024 Nuclear Energy Institute 3 Industry Adoption of AI The industry takes a judicious approach to new technologies, reflecting our conservative safety culture That approach applies to AI and includes:
Market and feasibility studies Detailed evaluation of concept: benefits, costs, risks, existing requirements and effects Regulatory interactions if necessary Proof-of-concept and pilot testing, including initial uses in business and non-safety applications Review of lessons learned and sharing of operating experience Adaptation to the business Further deployment/scale-up if desired/justified
©2024 Nuclear Energy Institute 4
©2024 Nuclear Energy Institute 4 NEI Member Interests in Artificial Intelligence AI Use Case Using Developing Future Interest Work Order Screening XXX XXX XXXX Work Scheduling and Optimization X
XXXXXXXX Work Package Development XXXXXXXXX Identifying Applicable Operating Experience XXX XXXXXXXX Identifying Training and Qualification for Work Packages XXXXXXXX Work Planning X
XXXXXXXX Parts Staging/Stocking X
X XXXX Inventory Management X
X XXXXX Supply Chain/ Logistics Optimization X
XXXXXXX
©2024 Nuclear Energy Institute 5
©2024 Nuclear Energy Institute 5 NEI Member Interests in Artificial Intelligence AI Use Case Using Developing Future Interest Condition Report - Initial Screening XXXXX XX XXX CR Prioritization and Scheduling XX XXX XXXXX CR Response X
XXXXXXXX Observations X
XX XXXXXXX Trending XX XXXX XXXX Preventive Maintenance XXX X
XXXXXX Corrective Maintenance X
XXXXXXXX Equipment/ Performance Monitoring XXXXX XX XXXX Nondestructive Evaluation XX XXXXXX Operator Decision-Making Aids X
XXXXXXXX Chatbot to Answer Questions XX XXXXXX XXX Operability Determinations XXXXXXXXX Reportability Determinations X
XXXXXXXXX Simulator/ Simulation Enhancements X
X XXXXXX Autonomous Operations and Control XXXXXXX Fire Watch X
XXXXXXX
©2024 Nuclear Energy Institute 6
©2024 Nuclear Energy Institute 6 NEI Member Interests in Artificial Intelligence AI Use Case Using Developing Future Interest System/ Component Modeling and Design X
XXXXXXX Design Optimization X
XXXXXXX Risk Assessment X
X XXXXXXX Maintenance Rule Determination/ Evaluations XXX XXXXXXX Core Design X
X XXX Reload Optimization X
X XXXXX Outage Planning and Management XXXXXXXXX Outage Optimization X
XXXXXXXXX Capture and Maintenance of Licensing Basis X
XXXXXXXXXX Research and Document Retrieval X
XXXXXXX XXXX Processing of Incoming Regulatory Documents X
XXXXXXXXX Development of Outbound Regulatory Submittals XX XXXXXXXXX Preparation of Mandatory Regulatory Reports XXXX XXXXXXXX
©2024 Nuclear Energy Institute 7
©2024 Nuclear Energy Institute 7 Recommendations Recognize that AI is evolving more rapidly than can most regulatory tools (rulemaking, regulatory guides, NUREGs)
Rely on the existing regulatory framework as much as possible Pursue a highly collaborative approach in the revision or creation of requirements for AI (e.g., Reactor Oversight Process),
including table tops and pilot tests Strive to keep the regulatory touch as light and flexible as possible to allow for continued rapid innovation Allow room and credit for industry-led, industry-developed guidance Resist succumbing to loud purveyors of the Terminator scenario
©2024 Nuclear Energy Institute 8
©2024 Nuclear Energy Institute 8 Final Thoughts NRC and industry must work together to educate the public about AI use in nuclear facilities and to maintain public trust.
Maintaining public trust may require higher levels of transparency about models, methods and applications than NRC and industry are accustomed to.
We appreciate the NRCs collaborative efforts thus far.
We look forward to working with the NRC to find an appropriate balance between regulatory oversight and creative freedom to enable NRC and industry to innovate with AI tools to improve safety, efficiency, and quality of work life.