Regulatory Affairs news highlights: June 2026

 

Recent regulatory headlines we’re tracking include:

 

U.S. House Energy Committee holds hearing on transmission permitting

The House Committee on Energy and Commerce held a hearing on “Wires, Rates, and States: Permitting Transmission for Reliable and Affordable Power” on May 13. The hearing witnesses were: Tony Clark, Executive Director of NARUC; Mark Christie, former FERC Chairman and Director of the Center on Energy and Law at William & Mary Law School; Randy S. Howard, General Manager of Northern California Power Agency; Clay Rikard, Senior Vice President of System Planning at Southern Company; Rob Gramlich, President of Grid Strategies; and Michael Skelly, CEO and Co-founder of GridUnited.

Mr. Clark testified that transmission planning should work in harmony with state policies and laws, and that states are generally supportive of federal permitting reform efforts provided they do not preempt state authority. Mr. Christie cautioned against giving FERC backstop siting authority, stating that FERC does not have the local expertise or bandwidth to evaluate and approve hundreds of transmission projects across the country. Mr. Rickard noted that state regulators are best positioned to weigh factors that are local in nature. Mr. Gramlich stressed the importance of transmission capacity, stating that interconnection between systems can achieve reliability at lower costs by sending power to the area that needs it most. Mr. Skelly also discussed transmission benefits, highlighting cost savings, reliability benefits from sharing reserves; and resilience benefits from providing alternate pathways during extreme weather or equipment failures. A full recording of the hearing is available here.

 

Trump Administration executive order on advanced AI directs enhancements to AI cyber security efforts

The Trump Administration issued an executive order, “Promoting Advanced Artificial Intelligence Innovation and Security,” that is meant to shore up security around the release of new Artificial Intelligence (AI) models. The executive order gives the secretary of the Department of Homeland Security (DHS) 30 days to begin standing up cyber security efforts for advanced AI at civilian federal government information systems and to facilitate access to tools and services for state and local governments and operators of critical infrastructure, including utilities.

The director of the Cybersecurity and Infrastructure Security Agency (CISA) and several cabinet secretaries will form an “AI cybersecurity clearinghouse, in voluntary collaboration with the AI industry and operators of critical infrastructure.” The effort will coordinate and deconflict scanning for software vulnerabilities and coordinate and prioritize how to remediate and patch any issues detected. The order also gives the White House and executive agency leaders 60 days to develop and maintain a classified benchmarking process to assess advanced cyber capabilities of AI models and to determine whether they should be labeled “covered frontier models.” The order creates a voluntary framework for firms that develop those models so they can provide the federal government with access to the models 30 days before they are released to address issues they could cause with the cyber security of critical infrastructure.

 

CISA, partner agencies issue guide on use of agentic AI services

CISA, the NSA, the Australian Cyber Security Center, and other international partners have issued a new guide titled “Careful Adoption of Agentic AI Services.” The guide defines agentic AI as being “composed of one or more agents that fundamentally rely on an AI model, such as an LLM, to interpret and reason about the state of the world, make decisions and take actions.”

The guide discusses risks and vulnerabilities associated with agentic AI. It also provides guidance on how to safely design, develop, deploy, and operate agentic AI systems and conduct risk assessments and mitigation activities. Key recommendations include: 1) using the principle of least privilege, 2) building security into the system architecture, 3) human oversight mechanisms and strong transparency practices, 4) inputting management controls, and 5) red teaming to assess the security and resilience of AI agents.