Article

A discussion among experts in power grid security focused on the potential use of AI within systems monitoring the North American Bulk Electric System (BES). Current Cybersecurity Incident Response (CIP) standards do not address AI, creating concerns about new risks.

Examples of AI failures, such as financial fraud and autonomous vehicle accidents, underscore the unpredictability of AI behavior. Generative AI models are complex black boxes that complicate risk assessment, making it difficult to ensure safe deployment. While AI can be used to provide advice, over-reliance without human oversight could pose serious risks to grid operations.