New York regulators review utility AI use cases

That shift is important economically because utilities are leaning on AI to reduce operating costs, predict outages, inspect infrastructure and handle customer service at a time when reliability demands and capital spending are both rising. New York’s regulators are not rejecting the technology. They are acknowledging the upside while warning about the downside: hallucinations, bias, data-privacy problems, misconfiguration errors, cybersecurity attacks and “functional brittleness.” In a sector where a bad decision can hit safety or service quality, even a small increase in oversight can reshape adoption timelines and force utilities to spend more on controls, audits and human review.

Consolidated Edison, which serves New York City, said it already uses AI for customer service, equipment inspection and mapping. National Grid has deployed GridCARE to help free up interconnection capacity for large-load customers, while the New York Power Authority has used AI to analyze drone data for vegetation management. Those are exactly the kinds of use cases the market should expect to proliferate across U.S. utilities as data centers, electrification and grid congestion drive a bigger need for automation. But New York’s move suggests the next phase of AI in utilities will be about governance as much as efficiency.

That has direct investment implications. The market often treats utilities as defensive dividend names, but the real opportunity may be in the tools, software and services that make AI safe enough to deploy inside regulated infrastructure. Cybersecurity, grid analytics, asset-inspection software, and compliance platforms stand to benefit if regulators elsewhere follow New York’s lead. Meanwhile, utilities with the best digital controls and cleanest governance frameworks can turn oversight into an advantage by proving they can deploy AI without compromising reliability.

The stock action already shows how investors are trying to sort winners from losers in a more demanding utility environment. Con Edison has held up better than some peers even as the broader utility tape has weakened, while other regulated names remain under pressure from rates, reliability spending and scrutiny over execution. New York’s audit raises the bar again, and that usually favors operators with scale, stronger technology budgets and tighter risk management.

My view: this is not a reason to avoid utility AI. It is a reason to own the picks and shovels around it and favor regulated utilities that can prove they have the controls to use it responsibly. The real money will be made in the next layer of the stack — the software, cybersecurity and infrastructure vendors that help the grid become smarter without becoming more fragile.