Attack the time tax that makes government such a burden for so many New Yorkers.
New Yorkers spend an astonishing amount of time waiting on the government.
We listen to hold music, scour our email (or dust off old filing cabinets) hunting for documents and click through websites that look like they were built 20 years ago. We fill out forms with information the government already has, again and again. And we call to check on an application because there’s no easy way to see its status online. When something goes wrong, we begrudgingly start the process again.
In a 2021 article for The Atlantic, journalist Annie Lowrey gave this phenomenon a name: the “time tax.” Her new book, “The Time Tax: How the Government Wastes Our Time — and How to Fix It,” expands on her argument. Policymakers should pay attention. Government red tape and outdated operations impose costs in hours of effort and frustration — costs that are harder for some New Yorkers to absorb than others. Someone with a flexible job, access to Wi-Fi and a computer, or with the resources to hire professional help has an easier time navigating this bureaucracy than someone working two shifts, caring for family, managing a disability or struggling to understand English.
Fortunately, AI gives New York City the ability to attack this problem — streamlining paperwork and applications and improving operational efficiency. Unfortunately, today, the mere mention of AI can turn a conversation about modernizing government services completely sideways. We feel like we’re forced to take sides (are you an accelerationist or a doomer?). And the mere suggestion that AI might be able to, for instance, help streamline NYCHA repairs could feel tone-deaf when most headlines we see are focused on existential risk.
Indeed, at the City Council’s headline-grabbing all-day October hearing on AI, much of the attention went to frontier models, whistleblowers and the possibility of catastrophic harm. Council members pressed AI companies to quantify the odds of a worst-case catastrophe, while some testified about scenarios in which humanity could lose control of advanced systems.
These are serious questions, and I don’t blame policymakers for asking them. But the answers will ultimately depend on national and international policy.
At the same time, New York City government has a responsibility and an opportunity to figure out how this technology can make everyday life easier for the 8 million people it serves — which should mean fixing the often broken interface through which New Yorkers try and fail to interact with their government.
A positive, local approach to AI might not generate headlines. But it could give a mother back an afternoon she would otherwise spend trying to recertify a benefit, or it could help an immigrant find a library program without waiting for an interpreter. Multiply those gains across millions of interactions with government and the organizations that provide public services, and AI could finally start to reduce the time tax that has become a scourge of overburdened New Yorkers.
City Comptroller Mark Levine recently touched on this when he suggested that AI could eventually allow someone to apply for SNAP simply by speaking into a phone in the language they are most comfortable using. His larger point was that, amid justified anxiety about AI’s biggest risks, New York cannot ignore the enormous amount of useful work sitting right in front of us.
A positive, local approach to AI could give a mother back an afternoon she would otherwise spend trying to recertify a benefit, or it could help an immigrant find a library program without waiting for an interpreter.
Anyone who has spent 20 minutes arguing with an automated customer-service agent understands why “AI-powered service” can sound like a headache waiting to happen. The Adams administration built an AI chatbot that was so clunky, it quickly became the butt of jokes.
I’m not talking about building a better version of that tool. Rather, AI can make government many times more efficient mostly by working in the background, helping staff pull relevant information from thousands of pages of records, identifying missing information before an application gets rejected and reducing repetitive data entry. Saving time now spent shuffling papers will give government employees more time to actually solve problems that require human judgment.
New York’s nonprofit sector already gives us an idea of how this might work.
Through Tech:NYC’s Decoded Futures program, we have helped hundreds of nonprofits in New York City and thousands more beyond the five boroughs explore how AI can ease the everyday operational burdens that consume staff time and limit how many people they can serve. These organizations don’t come to us asking how to replace employees. They come with the opposite problem: that their employees are overwhelmed by administrative work that keeps them from doing the jobs they signed up for.
Welcome to Chinatown, which supports small businesses in the neighborhood, offers a particularly good example of how a nonprofit can use AI to cut administrative burdens, freeing up more time for its core mission. Many small businesses and merchants in Chinatown have limited digital footprints and face language barriers, making it difficult for the nonprofit’s small staff to track which storefronts were opening or closing and determine the most efficient way to support business owners.
Staff and volunteers would walk the neighborhood with clipboards, then manually enter what they learned and spend a half hour preparing a single outreach email. Working through Decoded Futures, they built two custom AI-powered tools. The first was a volunteer canvassing app, which collects and updates data on nearly 1,000 storefronts, cutting a 10-hour task to about two hours. The second was a custom GPT that recommends what kind of support a business might need based on factors like how long they’ve been operating, reducing prep time for individualized business outreach from roughly 30 minutes to five.
Those saved hours can be spent sitting down face-to-face with merchants, building relationships and doing the work that actually requires a human touch.
We’ve seen similar successes across a range of organizations. Brooklyn Public Library developed a prototype that can help patrons find programs and resources in their preferred language instead of waiting for an interpreter. HeartShare, which serves New Yorkers with developmental disabilities and families facing instability, developed a tool that helps front-line staff parse dense records and quickly find critical information about an individual’s needs and preferences. In both cases, technology is reducing the administrative layer, speeding up processes and cutting down on a time tax — not replacing human labor or judgment.
City government should approach AI with the same mindset. Luckily, there are encouraging signs. Mayor Mamdani’s Commission on Government Efficiency has focused on the delays and procedural burdens that make government difficult to navigate, including contracting processes that oftentimes take months. His administration has also created Public Interest Technology (PIT) Crews, small technical builder teams that work with agencies to scale and deploy better digital services.
Every time government makes someone’s life easier or reduces friction in some way, it gives that person a small reason to believe that public institutions can work.
The next step should be to identify the public processes that consume the most unnecessary time, beginning with those used disproportionately by low-income residents and people who depend on government benefits. Agencies should systematically measure how long their interactions with New Yorkers typically last, how often people have to resubmit information or ask for help and where online applications get abandoned.
For example, CityFHEPS, the rental assistance program that helps New Yorkers find housing, has long underperformed and been plagued with delays. The City Council identified missing paperwork, errors and the manual sorting of uploaded documents as sources of friction. A targeted AI approach could help staff organize documents into the right case files, pull relevant information from leases and pay stubs, and flag missing items early, giving caseworkers more time to help families resolve problems and get them into housing faster.
These issues extend to all aspects of government, including the simple act of visiting a city benefits office. In fiscal year 2026, New Yorkers waited an average of 97 minutes for in-person service at the City's Human Resources Administration offices, up from about 44 minutes in 2022, according to the Mayor’s Management Report. This is the time tax at its core — the people who need the most support are often the ones with the biggest hurdles in their way. AI could help residents identify and fix problems with benefit applications before they require a trip to an office. For those who do need to visit, it could prepare a summary of their case for staff so they don’t have to explain their situation from scratch.
At every juncture, government workers should also ask a difficult question: Will speeding up or automating a certain process help New Yorkers, or is there an underlying issue rooted in bad policy? Sometimes a process is difficult because the rules are outdated or unnecessarily complicated — and automating a bad process could backfire. AI can improve service, but when used carelessly, it can exacerbate the very pain points we are trying to eliminate.
New Yorkers already know what good public technology feels like. Think about walking into the subway and tapping a phone or credit card at an OMNY reader. There is almost nothing to think about — you tap and keep moving. That kind of interaction may seem mundane, but the competence matters enormously. Every time government makes someone’s life easier or reduces friction in some way, it gives that person a small reason to believe that public institutions can work. As New York City Chief Technology Officer Lisa Gelobter put it when the PIT Crews launched, “Every confusing process, unnecessary step or difficult interaction makes government harder to navigate. Every simpler experience strengthens trust.”
The biggest believers in AI often point to its potential for scientific breakthroughs and medical discoveries, and those bold predictions get lots of attention and excitement. But giving New Yorkers more and more of their valuable time back by making our city easier to live in could be just as impactful — and we don’t have to wait years to see the results.





