Hochul Says New York Is Using AI to Audit Old Rules as State Tightens Data Center Oversight

New York Governor Kathy Hochul says her administration used artificial intelligence to review the state’s laws, regulations and policy guidance in a matter of months, a task she says would have taken years if done manually. The move underscores how quickly AI is moving from a technology-policy problem to a government tool, even as New York simultaneously imposes a pause on new hyperscale data centers.

Speaking on Bloomberg’s Odd Lots podcast, Hochul said the state used AI to identify outdated or contradictory rules it could remove or modernize, framing the effort as part of her broader push to make government faster and less burdensome for residents and businesses.

The governor’s comments land at a politically sensitive moment. Just days earlier, New York became the first state to approve a moratorium on new hyperscale data centers for up to a year, reflecting concern over electricity demand, water use and rising utility costs. Together, the two developments show the state trying to balance enthusiasm for AI’s administrative uses with caution about the infrastructure that powers it.

What Hochul says New York used AI to do

Hochul said her team relied on AI to scan “every single rule, regulation and policy” on the books in order to find outdated provisions that no longer make sense in modern life or government operations. She said the process surfaced old rules that could be eliminated without harming public safety or essential oversight.

Among the examples she cited were a $25 fee tied to hunting with dogs and a requirement that pregnant people obtain a permit to work after midnight. Those examples were presented as evidence that the state’s regulatory code still contains provisions that are widely considered obsolete, confusing or out of step with current expectations.

In Hochul’s telling, the technology helped compress what she suggested could have been a multiyear bureaucracy project into a short-term cleanup effort. Rather than relying only on staffers to manually read through vast archives of law and administrative rules, the state used AI to surface candidate reforms more quickly.

“Probably would have taken five years at the staff level,” Hochul said of the review, adding that AI allowed the work to be done “in a couple of months.”

The governor said the point was not to replace human decision-making, but to help state officials identify which regulations are outdated and can be removed or revised. The final decision, she emphasized, remains in the hands of government leaders and agencies.

Why does this matter now?

This matters because New York is positioning itself as both a regulator and a user of AI at the same time. That dual role is increasingly common among governments, but Hochul’s comments make the contradiction unusually visible: the state is making it harder to build more AI infrastructure while using AI internally to streamline government.

That tension is especially relevant in New York because the state is one of the country’s biggest economic and political hubs. Any shift in how it writes, interprets or removes regulations can have an outsized impact on businesses, workers and residents.

Hochul’s argument is essentially that AI can help government become more responsive and less cluttered. Her critics are likely to focus on whether the process was transparent, how the model was supervised, and whether an algorithm can reliably distinguish between truly obsolete rules and laws that still serve an important purpose.

Those questions go beyond New York. Other states and federal agencies are also exploring where AI can help with document review, customer service, compliance work and internal operations. New York’s example may become a reference point for public-sector adoption, particularly if officials disclose how the system was tested and what safeguards were used.

How did New York get here?

New York’s AI-assisted rule review is happening alongside a broader backlash against the physical footprint of AI infrastructure. The rapid growth of data centers has raised alarms in states across the U.S. over energy demand, strain on local grids and the use of water for cooling systems.

Earlier this week, New York enacted the first state-level pause on new hyperscale data centers, a temporary moratorium that can last for up to a year. Lawmakers are expected to use that time to draft rules aimed at protecting ratepayers and natural resources from the impact of massive computing facilities.

The state’s approach reflects a growing policy split: officials may be eager to use AI tools for administrative efficiency, but they are becoming more skeptical about the costs of building the infrastructure that supports large-scale AI systems.

In practical terms, a hyperscale data center is a massive computing site that can house thousands of servers and consume large amounts of electricity. These facilities are central to cloud services and advanced AI systems, but they also raise questions about who pays for the additional grid capacity and environmental load.

What the data center pause signals

The data center moratorium suggests New York is trying to get ahead of a problem many states are only beginning to confront. AI systems need enormous computational power, and that demand can translate into higher energy use, more transmission pressure and additional water consumption.

By pausing new hyperscale builds, New York is effectively asking whether the public gets enough economic benefit to justify those trade-offs. The answer could shape how other states design incentives, permitting processes and utility rules for AI-era infrastructure.

How much can AI help government cleanup work?

AI can accelerate large-scale document review, but it is best understood as a filter rather than a final authority. In government settings, the technology can help sort through huge numbers of records, identify likely duplicates, and flag provisions that may deserve human review.

That is especially useful for states with dense legal codes that have accumulated amendments, exceptions and legacy language over decades. Rules can survive long after the real-world conditions that created them have disappeared, and staff time alone often limits the pace of review.

Still, the benefits come with caveats:

• AI can misread context or miss subtle legal distinctions.
• Old laws may appear outdated but still support a current policy objective.
• Automated review requires strong human oversight and clear records.
• Public trust depends on whether the process is transparent and explainable.

Hochul’s message suggests that New York used AI as a productivity tool, not a replacement for legal judgment. But the broader debate over automated governance will likely hinge on how much discretion officials are willing to hand over to software, even indirectly.

What kinds of laws did Hochul highlight?

Hochul used a few colorful examples to illustrate the age and oddity of some state regulations. The point was not just that the laws were outdated, but that they revealed the sheer volume of rules that can survive in a system built over many decades.

These examples are striking because they help the public understand why governments struggle to modernize without new tools. Even when a law appears trivial, it still requires legal review, administrative action and sometimes legislative approval to remove.

By using memorable examples, Hochul is also making a political case: that AI can help elected officials show visible results quickly. For governors, speed matters because regulatory cleanup is often invisible to voters unless it is tied to a concrete story about waste, inconvenience or outdated bureaucracy.

Why is this politically useful for Hochul?

Hochul is presenting herself as a technologist-friendly governor who wants to use AI where it clearly improves public administration. At the same time, she is drawing a line between using AI as a back-office tool and allowing the AI industry to expand without limits.

That message may appeal to two different audiences. Business leaders may like the idea of a government that is easier to deal with, while residents may welcome a state that removes outdated legal clutter. Meanwhile, environmental advocates and consumer groups may see the data center pause as evidence that the administration is at least willing to regulate AI’s physical footprint.

Her framing also gives her a rhetorical advantage. Rather than treating AI as a futuristic abstraction, she is describing it as a practical tool that can help cut red tape. That kind of example is easier for voters to grasp than technical conversations about models, GPUs or training data.

Hochul said she wants “a government that’s not on your back but on your side,” arguing that AI has helped her administration move faster and make “dramatic changes.”

Whether that message resonates will depend partly on whether New Yorkers actually see changes in the form of simpler rules, faster service and fewer outdated requirements. If the cleanup is real and visible, the governor can point to AI as an efficiency success. If not, the effort risks being seen as a symbolic proof of concept.

How other governments may read New York’s example

Other states and municipalities are likely watching closely because New York has now shown both sides of the AI policy equation. It is using AI internally to improve administration while also acting as a regulator of the AI industry’s physical expansion.

That combination may become a template for governments that want to capture AI’s productivity benefits without letting the technology reshape local utilities, zoning and environmental planning on industry terms alone.

Possible lessons for public agencies

1. AI can speed up document review, but only with human oversight.
2. Old regulatory codes are often more complex than agencies can easily audit manually.
3. Infrastructure policy and software adoption must be treated as linked issues.
4. Public disclosure will matter if officials want trust in AI-assisted governance.

If New York can show that AI helped identify obsolete laws without introducing errors or confusion, it may encourage other governments to try similar tools. If the process produces mistakes, critics will use that as evidence that government still needs to move more cautiously.

What happens next in New York?

The immediate next step is likely a combination of administrative cleanup and legislative work. AI may help identify what should be removed or rewritten, but lawmakers and agencies still have to decide which changes actually move forward.

At the same time, state officials will begin work on the policy framework for the data center pause. That process will likely involve utility regulators, environmental concerns, economic development priorities and technology stakeholders that want more clarity on what kinds of projects New York will permit after the moratorium ends.

The governor’s comments suggest she sees AI as a lever for reform rather than a threat to be avoided. Yet the state’s data center pause shows that support for AI does not automatically extend to every part of the AI ecosystem.

In that sense, New York is becoming a test case for a more mature AI policy model: one that welcomes the software side of artificial intelligence when it improves government, but scrutinizes the power-hungry infrastructure that makes modern AI possible.

Bottom line

Hochul’s remarks are significant because they place New York at the center of a national debate over how governments should use AI. The state is leaning on the technology to modernize its own bureaucracy while also taking a hard look at the costs of large-scale AI infrastructure.

That dual strategy could shape how other states approach the technology: not as a single policy issue, but as a bundle of questions about administration, energy, environment, transparency and public trust.

For now, Hochul is betting that AI can help New York clear away outdated rules faster than human staff alone could manage. The bigger question is whether the state can do that while still keeping the process accountable, understandable and genuinely in the public interest.