At Norm Ai’s inaugural Central Park AI Forum, leaders from government and industry gathered to examine how AI may impact the economy and government. Two conversations highlighted how government officials are considering AI’s benefits: one focused on financial regulation in the age of AI, the other on building the physical and policy infrastructure that will enable AI to reach its potential.

NY Financial Services Regulation, Before and After Advanced AI

Adrienne Harris, Superintendent of the New York Department of Financial Services (NYDFS)

Ben Lawsky, Former Superintendent of NYDFS; Norm Ai Legal AI Committee

In a conversation between New York’s current and former top financial regulators, Superintendent Harris and former Superintendent Lawsky reflected on how advanced AI is reshaping both financial institutions and the agencies that oversee them.

Superintendent Harris described a financial sector that is adopting AI deliberately. Although insurers and banks are using AI for claims management, customer service, and underwriting, banks have proceeded somewhat more cautiously in these use cases. Across both banking and insurance, she said, the emphasis is on solving specific operational problems rather than adopting AI for its own sake.

Superintendent Harris argued that effective regulation should evolve without abandoning foundational principles. She discussed how existing standards, such as the prohibition on unfair discrimination in insurance underwriting, continue to apply in an AI context. For regulators, that means requiring firms to test their models, establish governance frameworks, and maintain accountability for any third-party technology they use. As Harris put it, institutions cannot outsource responsibility for model outputs.

The conversation also looked ahead to how AI could change supervisory and compliance processes. Superintendent Harris noted that machine-readable and machine-executable data and processes could allow regulators and firms to exchange data automatically, enabling more efficient examinations, also noting that humans in the loop will be key to maintaining quality for these processes.

The discussion turned to the state’s role in digital-asset oversight, an area Lawsky pioneered with the BitLicense framework. Harris described how that early groundwork positioned New York to regulate emerging technologies responsibly, noting that recent federal proposals on stablecoins and market structure (GENIUS Act) closely align with the NYDFS’s guidance, particularly around one-to-one reserves and timely redemption standards.Cybersecurity emerged as another area where AI is changing the landscape. The Superintendents discussed the NYDFS’s recently updated Part 500 cybersecurity rule, which strengthens governance while acknowledging that AI introduces both new risks and new defenses. As Harris noted, AI can heighten cyber threats by enabling much more sophisticated attacks, but it also enhances detection and response.

The speakers also addressed how regulators are beginning to use AI internally. NYDFS is investing in a multiyear, multimillion dollar technology transformation that includes AI tools for public inquiries, automated review of company filings, and a partnership with the University at Albany to combine weather, insurance, and mortgage data for analyzing climate-related risk.

The conversation ended with a forward look at how AI could modernize supervision itself, from the regulators’ and regulated firms’ AIs conducting portions of examinations autonomously, to real-time data sharing mitigating delayed insights seen in recent banking crises.

The Infrastructure for AI: Building AI's Backbone

Senator Mike Lee, U.S. Senator

Lane Dilg, Former Header of Infrastructure Policy & Partnerships at OpenAI

Will Kinzel, Global Head of Government Affairs at Carlyle

The closing panel of the Forum turned to what must be built quickly for AI to reach its full potential. Moderator Will Kinzel framed it as the practical discussion after a day of vision: how will we assemble the physical and policy assets that drive AI’s growth?

Lane Dilg opened by describing the U.S. position as a “sprint with hurdles.” The country leads in innovation and frontier models, she said, but faces clear obstacles, chief among them energy. AI’s future, she added, depends on overcoming those constraints and accelerating technologies such as fusion, where AI can already help shorten scientific discovery timelines.

Senator Mike Lee approached the issue from first principles: the success of AI depends on the inputs, the machine, and the outputs. Chips and data centers form the machinery, but without stable, baseload electricity, he warned, the system cannot scale. Lee noted that while the U.S. has made progress on chips, its energy growth lags China’s, framing the next five years as crucial for maintaining competitiveness.

The mix of energy sources likely to power AI was central to the conversation. Panelists cited natural gas as the near-term bridge until advanced nuclear and storage solutions can be deployed at scale. Senator Lee pointed to small modular reactors and geothermal projects as promising but years away from broad implementation. Dilg added that AI can accelerate the very science needed to overcome these constraints, from fusion research to materials discovery for energy storage, but those advances arrive after this crucial window.

Before turning to historical lessons, Senator Lee reflected on AI’s broader promise, as a technology that could enable people to live longer, healthier, and more productive lives. Beyond its infrastructure demands, he suggested, the true measure of success will be how effectively AI improves everyday work and human well-being. From there, Senator Lee broadened the discussion to lessons from past technological revolutions. He compared AI’s infrastructure challenges to moments when the U.S. led in transformative industries (railroads, aviation, the internet), arguing that success came from innovation driven by the private sector, not the government. As an example, he cited the Wright Brothers, who built the first working airplane with their own resources while a well-funded government program failed to achieve flight.

Dilg agreed that private capital must lead, touching on recent related financial innovations such as GPU-backed loans, layered financing structures, and other innovative models bringing investment into AI infrastructure at scale. She and Senator Lee acknowledged that projects like OpenAI’s Stargate, which require hundreds of billions in capital, naturally concentrate participation among a few players. Both cautioned that large government funding packages could deepen that concentration. They agreed that the government’s most effective contribution is to remove obstacles rather than fund projects directly, streamlining permitting, reducing delays, and cutting through the layers of regulatory apparatus that slow down development.

In closing reflections, Senator Lee emphasized that the opportunity of AI is vast but the risk of falling behind China and other countries is high, and now is the time to act to accelerate AI in the United States.