New York Orders Utilities to Disclose All AI Use Within 60 Days

New York regulators demand AI transparency from utilities

New York's Public Service Commission has ordered the state's utilities to disclose all artificial intelligence systems they currently use in operations, giving companies including National Grid 60 days to comply with the sweeping inquiry launched Thursday.

The directive requires utilities to reveal not only where they've deployed AI but also the policies, procedures, and protocols governing the technology's use across their organizations.

Why it matters

Utilities control critical infrastructure that millions depend on daily for power, gas, and water. As these companies quietly integrate AI into core functions—from answering customer calls to inspecting facilities—regulators are playing catch-up on understanding the risks. This inquiry represents one of the first comprehensive attempts by a state regulator to map AI use across an entire utility sector, potentially setting a template for oversight nationwide.

AI now embedded in utility operations

According to the commission, utilities have already integrated AI systems into fundamental operations. The technology now handles customer service interactions, performs predictive maintenance on equipment, and conducts facility safety inspections—functions that directly affect service reliability and public safety.

The extent of this deployment appears to have prompted regulatory concern about oversight gaps.

Regulators cite multiple risk categories

The Public Service Commission outlined a comprehensive list of AI-related vulnerabilities that motivated the inquiry. These include:

• Hallucinations: AI systems generating false or misleading information
• Algorithmic biases: Systematic errors that could disadvantage certain customer groups
• Transparency issues: Difficulty understanding how AI systems reach decisions
• Data privacy concerns: Risks to customer information processed by AI
• Misconfiguration errors: Technical mistakes in AI system setup
• Cybersecurity attacks: Vulnerabilities that hackers could exploit
• Functional brittleness: AI systems failing unexpectedly under certain conditions

The commission's framing suggests regulators view AI as introducing novel risks that existing utility oversight frameworks may not adequately address.

Next steps for oversight

Once utilities submit their disclosures, the Public Service Commission will evaluate whether current safeguards are sufficient or if new protections and policies need to be implemented. The 60-day timeline suggests regulators want rapid visibility into AI deployment before determining what additional rules may be necessary.

The inquiry could lead to formal requirements around AI testing, documentation, human oversight, or restrictions on AI use in certain critical functions.

Details of the order were first reported by WWNY.