Appendix C: RF Reliability Committee Feedback
RF’s Reliability Committee identified Artificial Intelligence risks as an additional area that is likely to impact risk assessments into the future. The committee’s input is summarized below.
AI introduces both operational opportunities and reliability risks when deployed in the electric sector. From a risk perspective, AI can amplify cyber threats by enabling faster, more scalable, and more adaptive attacks, including automated vulnerability discovery, phishing, malware development, reconnaissance, and data exfiltration. As reflected in the Security section, one emerging concern is the rise of AI-orchestrated attacks, where adversaries use AI to perform a large share of tactical cyber activity. In addition, expanding the use of AI-enabled digital systems may deepen reliance on third-party software, cloud services, communications systems, and vendor-managed tools, which can widen the attack surface and create additional vulnerabilities.
AI also creates indirect reliability risks through rapid growth in computational load, especially from data centers supporting AI model training and deployment. As described in the Grid Transformation and Modeling sections, this growth can increase uncertainty in demand forecasting, resource adequacy, transmission planning, and interconnection studies. If not modeled accurately, transmission and operational-based studies may not identify reliability concerns driven from large computational loads.
As reflected in the Situational Awareness section, ineffective use of Operating Instructions can create a direct reliability risk because operational commands must be clear, concise, and timely. Communication breakdowns can reduce situational awareness, impair coordination across control centers and other critical infrastructure sectors, and increase the likelihood that operators respond incorrectly or too slowly during stressed system conditions. These concerns are heightened during complex events, where reliable communication between electric, natural gas, water, transportation, and communications partners is essential to maintaining grid stability. With the addition of AI technologies like deepfakes, voice cloning or fabricated video could be used to impersonate trusted leaders, operators, or partner organizations, potentially resulting in false operating instructions, fraudulent requests, or delayed response while authenticity is being verified. This can undermine confidence in critical communications, increase the chance of misinterpretation or inappropriate action, and erode trust in digital evidence and coordination channels that are essential for maintaining situational awareness and grid stability.