New York Governor Kathy Hochul said her team used AI to analyze every rule, regulation, and policy across New York State, completing in "a couple of months" a review she estimated would have taken five years at the staff level. Hochul disclosed the initiative during a July 15, 2026 appearance on Bloomberg's Odd Lots podcast, hosted by Joe Weisenthal and Tracy Alloway (The Verge).

The AI-assisted review identified outdated legislation Hochul intends to eliminate. She cited two examples: a $25 fee required to take a dog hunting, and a requirement that pregnant people obtain a permit to work after midnight. Both date to eras long preceding current regulatory frameworks and had not been flagged through conventional review channels.

Hochul framed the effort as part of a broader push to reduce bureaucratic friction. "I want a government that's not on your back but on your side, and using AI has been powerful to do that," she said (The Verge). She also urged wider adoption: "I think every level of government should use this… I'm going to make dramatic changes using the power of AI."

The Odd Lots episode, titled "NY Governor Kathy Hochul on Her One Year Data Center Moratorium," covered more than the regulatory audit. Earlier the same week, New York became the first state to impose a moratorium on new hyperscale data centers, lasting up to one year, to allow the state to develop rules protecting the environment and energy grid from power-hungry AI facilities (AP News; The Verge). Bloomberg also published a related Odd Lots newsletter article on July 15 titled "The Optimistic Case for the Hochul Data Center Pause."

The regulatory review connects to a formal policy track Hochul launched earlier in July. On July 8, she issued an Executive Order commencing what her office described as a "Regulatory Reset," a comprehensive and methodical review of thousands of state regulations (governor.ny.gov). The AI analysis appears to serve as the technical engine for that initiative, giving agencies a structured inventory of rules to act on.

Hochul's AI policy agenda extends beyond deregulation. In March 2026, she launched the FutureWorks Commission to guide policy and private-sector interventions in response to the impacts of AI on workers across New York (governor.ny.gov). In January, she announced she wants to prohibit political campaigns from spreading AI-generated images of people (NYT on Facebook). Taken together, these moves sketch a policy posture that is simultaneously permissive about using AI inside government, cautious about its infrastructure footprint, and restrictive toward its potential for disinformation.

The broader context here is a striking juxtaposition. A governor who just paused hyperscale data center construction to protect the grid is simultaneously championing AI as a tool to overhaul the regulatory state. There is no inherent contradiction: the moratorium targets the compute-intensive training and inference infrastructure that strains power supply, while the regulatory audit is a text-analysis task that could plausibly run on modest infrastructure. The term "hyperscale" refers to data centers large enough to serve major cloud providers, often consuming as much electricity as a small city. The timing does illustrate how quickly AI has moved from an infrastructure question to an operational one inside state government.

The more substantive question is what "dramatic changes" will actually look like in practice. Hochul has said the AI analysis will allow her and state agencies to eliminate outdated regulations. What remains unspecified is the review and validation process before those eliminations take effect. An AI system can flag a statute as antiquated, but the legal mechanics of repealing a regulation involve public comment periods, agency review, and often legislative action. The five-year estimate for manual review likely accounts for those procedural steps; the two-month AI timeline almost certainly does not.

That gap between identification and action is where the practical work begins. A large language model can ingest the full corpus of New York's administrative code, flag statutes that reference obsolete technologies or outdated fee structures, and produce a ranked inventory for human review. That is genuinely useful and would have been prohibitively expensive at scale even three years ago. But it is a triage tool, not a repeal mechanism.

Hochul's framing positions AI as a way to make government leaner and more responsive. That is a defensible claim if the follow-through includes the procedural rigor that regulation demands. The examples she cited, a hunting-dog fee and a midnight-work permit for pregnant people, are the easy cases: clearly archaic, broadly uncontroversial, and politically costless to remove. The harder test will come when the AI surfaces regulations that have constituencies defending them, or where "outdated" is itself a contested judgment.

For state and local governments watching from the outside, the New York experiment offers a template worth tracking. The combination of an executive order mandating review, AI as the scanning mechanism, and a public commitment to elimination creates a feedback loop that could accelerate regulatory modernization. Whether it does depends entirely on what happens after the AI delivers its findings.