The AI bargain cities should strike with labor
Recently, Mayor Zohran Mamdani announced a Civil Service Modernization Advisory Committee that will study the outdated civil service system that hinders workers and frustrates residents. This important step could help him close the gap between his ambitions and budget realities.
Expenses continue to rise faster than revenues, a structural challenge with no easy answers. The mayor’s Office of Management and Budget projects budget gaps totaling more than $23 billion from fiscal year 2028 to 2030. That doesn’t even factor in funding for signature Mamdani priorities, such as universal childcare and free buses.
The usual response involves choosing between raising taxes and cutting services and, from time to time, resorting to temporary budget gymnastics or asking Albany for money. But with dramatically improved technologies, including AI and related data insights, city leaders now have a better option: transform how government works and deliver better services at lower cost.
AI represents the most significant opportunity to modernize urban governance in a century. If Mamdani and his team capture it, they could reset the fundamentals of public-service economics. AI changes what one worker can accomplish in a day. Technology can automate rote tasks, such as processing applications and answering routine questions. Public works teams can better predict maintenance needs, preventing potholes or streetlight outages before they happen. Technology can dramatically reduce the drudgery that defines so much public-sector work, replacing it with problem-solving that connects workers with residents’ needs rather than the often needless routines of outdated procedures.
Workers should welcome changes that free them from basic tasks and let them focus on problem-solving. Residents and businesses will feel the difference, too. Any New Yorker who has tried to open a restaurant — a process that can touch more than a dozen agencies and take the better part of a year — would gladly hand the paperwork to a City Hall AI agent trained and contained to handle that job.
If city government were a small business, it could use AI freely to forgo hiring a designer or a bookkeeper. It isn’t; its massive workforce has protections written into state law and public-sector union contracts. As a result, even if an AI innovation emerged that could do a particular bureaucrat’s work far faster, cheaper and better, it would be exceedingly difficult to implement.
That means realizing these gains will require creativity and adaptability from management and labor alike. Unfortunately, across the country, local labor negotiations have overwhelmingly focused on constraining AI rather than capitalizing on it. In a move meant to slow or stop AI deployment in cities, the American Federation of State, County and Municipal Employees (AFSCME) passed a resolution demanding that “every stage of the process, including design, implementation and monitoring of AI systems” become subject to collective bargaining.
This is understandable. Union leaders worry that technology will replace their members, hurting working people and their families. The worry is legitimate; the response is not. Blocking AI at the bargaining table will not permanently save jobs. It will simply ensure that, when budget pressure inevitably forces adoption, it happens on worse terms for workers, with no training, no productivity sharing and no protection against layoffs.
A smarter path forward is for city and union leaders to consider how technology and people can work together more effectively. Getting labor to the negotiating table in this spirit may take some incentives. In exchange for embracing responsible AI rather than resisting it, cities should offer workers three concrete commitments — commitments that could and should be part of Mayor Mamdani’s labor negotiations: a promise to forgo AI-related layoffs, an investment in training on the new tools and the conversion of a share of the productivity gains into workers’ paychecks.
This is not theoretical. During my time as mayor of Indianapolis in the 1990s, we made exactly that kind of agreement — and it worked. Facing a billion-dollar infrastructure deficit and higher property taxes than those in our surrounding suburbs, we opened municipal services to competitive bidding.
Essentially, we let private firms and city workers compete to deliver everything from filling potholes to printing forms. Freed to think creatively, incumbent workers offered hundreds of suggestions to cut costs and ultimately won many contracts. We paid them bonuses for productivity gains and pledged not to lay off union workers for identifying efficiencies. Good ideas shouldn’t cost you your job.
Over time, Indianapolis saved hundreds of millions of dollars through employee attrition. Thanks to an agreement with the unions to avoid layoffs, the non-public-safety local government workforce shrank by 40% over eight years through retirements and other voluntary departures. Empowered sanitation and fleet workers, with less stifling managerial oversight and incentivized with bonuses, helped redesign their work. The numbers may look different in a city as large as New York. Still, there’s a big opportunity to match attrition with productivity to help alleviate structural deficits.
New York City government today employs roughly 300,000 people and turns over about 10% of that workforce every year through retirements and voluntary departures. Declining to refill even half of those slots would produce billions in savings, counting pay, benefits and pensions that extend for decades. As a general rule of thumb, one $100,000 job costs the City nearly $150,000 a year once you account for pensions, health insurance and payroll taxes. Over a 25-year career plus retirement, it obligates taxpayers to pay close to $4 million.
The City doesn’t need waves of layoffs, but with the advent of advanced AI systems, there is no reason New Yorkers should accept that they need a government controlled by antiquated rules. Over time, government can and should be right-sized.
To fully unlock AI-driven gains, New York needs to modernize Progressive Era civil service rules. Job classifications box public employees into rigid roles and routines, often with multiple layers of supervision. If AI is to take over rote work so that humans can focus on harder tasks, city workers must be freed to exercise more discretion than current rules allow. A 21st-century technology layered atop a 19th-century personnel system will produce neither efficiency nor the meaningful public-service careers young people — inspired by Mamdani — seek.
Striking an AI bargain with labor, matching creative productivity with attrition and less supervision, will empower workers and give them a reason to adopt AI willingly. The technology is already here, and budget desperation will force its adoption on government sooner or later. Better to shape how this change happens than to watch it arrive haphazardly — imposed piecemeal by budget crises, vendor lock-in and public frustration, without the worker protections a bargain could secure. This is not only an opportunity to balance budgets. It is a chance to define what modern public service looks like, partner with the practitioners doing the work and increase responsiveness and trust in local government.





