This year, I broke the Yom Kippur fast the way generations of New Yorkers have—with bagels, lox, and—after checking my receipt—some thoughts about the price of food in this city.
I complained to my friends. But high prices at the Jewish dinner table are nothing new, and perhaps I should have looked to our forebears for some inspiration.
In the spring of 1902, the price of kosher meat on the Lower East Side suddenly jumped by 50 percent—from 12 to 18 cents a pound. Not because of a cattle shortage or drought. A small group of powerful meatpacking companies—the beef trust—had come to dominate the market.
For the Jewish immigrant families packed into the neighborhood’s tenements, that wasn’t just sticker shock. It meant putting less food on the table. But the women responsible for buying and preparing this food did not resign themselves to this price increase. Two women, Fanny Levy and Sarah Edelson, organized their neighbors along Monroe Street door to door. They picketed butcher shops. They took their campaigns into synagogues, climbing onto bimahs during Shabbat services and calling on their communities to join the boycott. They passed out flyers bearing a skull and crossbones and a simple message: “Eat no meat while the Trust is taking meat from the bones of your women and children.”
People listened. The protest grew to involve some 20,000 New Yorkers. Kosher meat sales collapsed. Within weeks, the retail price had fallen from 18 cents to 14 cents a pound.
The women had no control over the meatpacking industry. But they could see what it was doing. The price hike was public. They could compare notes, recognize a shared interest, and organize around a common demand.
Now imagine that same fight unfolding in 2026. Sarah Edelson opens a grocery delivery app and sees a price for meat. Fanny Levy opens the same app and sees another. The woman down the block sees a third.
The company knows what each woman has bought before, where she lives, and what alternatives are nearby. Perhaps it can estimate how urgently she needs to place an order—say, before the Sabbath. Each woman sees only the price offered to her.
Think about what that does to the boycott. Assuming these women even realize they’re being gouged, what price goes on the flyer? What number does Sarah shout from the bimah? How do you organize your neighbors against a price hike when you can’t tell whether they received one?
The beef trust had enormous power. But its power was exercised in a way that people could see and challenge. Today, companies have increasingly sophisticated tools to exercise economic power person by person—consumer by consumer, worker by worker.
The company sees everyone while each of us sees only ourselves. And that gets at what worries me most about how AI is reshaping our economy.
On October 5, representatives from the major AI firms will testify before the New York City Council to quell concerns about human extinction, to walk back their own promises of mass layoffs, and to give their best pitch on why, in Sam Altman’s words, artificial intelligence should be a “utility like electricity or water,” where “people buy it on a meter.”
These representatives will make the case for what this extraordinary technology can do. Our job at the Department of Consumer and Worker Protection (DCWP) is to ask how it affects the people who have to live with it. That is the work we have been leading since day one of the Mamdani administration—and, as I will outline here, the work we intend to keep doing in the months and years ahead.
I can already predict that one phenomenon the AI companies will focus on—especially if it distracts from their safety record—is the promise of personalization. And you can understand the appeal: Personalized products, personalized services, personalized experiences all sound good. But there is a darker side when companies personalize the terms on which people earn a living or buy the things they need. A price based on the most you are willing to pay. A wage based on the least it thinks you will accept.
Companies, largely unchecked by government, have spent years building the machinery for this: collecting data, tracking behavior, and testing our responses. AI can make that machinery more powerful and expand its reach. Firms can learn from millions of people while keeping each person in the dark about how others are being treated.
Agentic AI could take this individualization further still. These systems promise to do more than recommend a product: They could shop, negotiate, and make purchases on our behalf. That could save people time and help them find better deals. But it also means the products we encounter, the alternatives we consider, and the terms we accept could all be selected through a private exchange between automated systems that no human ever reviews.
That raises basic questions about whose interests these agents serve. Will an agent search for the best deal for its user, or steer them toward a business that pays for placement? What personal information will it reveal to a seller? And if it exposes our sensitive information or initiates an unwanted purchase, how can consumers stop it and get relief? As more of the transaction happens out of sight, consumers may find it harder to understand their own treatment, much less compare it with anyone else’s.
The threat AI poses to our ability to act collectively may be new. But our response doesn’t have to be. Fanny Levy and Sarah Edelson didn’t have the money or market power of the beef trust. But they had neighbors. They had shared interests. And once they recognized those interests, they discovered that acting together gave them power of their own.
I believe the same principle should guide us today, in New York City and around the country. If AI gives corporations powerful new ways to divide us, we need to build new ways—and strengthen some very old ones—to act together.
In New York City, we are working to reclaim an economy centered on the interests of working people. This means common rules that corporations cannot ignore. A minimum wage an algorithm cannot slip below. A fair price that doesn’t change based on who’s doing the shopping. Workers with the freedom to organize and bargain together. Governing institutions with the capacity and will to hold tech companies to account.
We are pursuing a three-part strategy for realizing this vision. We are enforcing the laws on the books, and firmly rejecting tech companies’ demands for special exemptions. We are defending our shared economy—common prices and common wages that working people can see and organize around. And we are investing in institutions of countervailing power—worker organizations, consumer advocates, and government itself.
Enforcing the Law
On September 22, DCWP announced the largest labor enforcement action in New York City history: a $131.5 million settlement with DoorDash, which admitted to underpaying its workers—and, in some cases, not paying them at all. The action will put more than $115 million back into the pockets of 260,000 workers.
But behind that headline is a deeper story. App companies have built a business model around extraordinary control. They can surveil workers in real time, measure their performance to squeeze extra output, adjust pay and incentives to keep enough people on the road, and keep vast reserves of workers waiting for work off the road.
In New York, we have insisted on a basic limit to that power: Delivery workers are entitled to minimum pay. Whatever a company’s system spits out, companies must still meet that obligation.
That is why this victory matters beyond the nine-figure recovery itself. It shows that a law enacted in this city can still set the terms of a market run through an app. An algorithm may calculate how little it thinks a worker will accept. But that algorithm doesn’t get the final say—the people of New York do.
Another development behind the headline is our cutting-edge approach to remedies. In this case, we are turning DoorDash’s massive worker surveillance apparatus on its head. In partnership with the Workers Justice Project and the Workers’ Algorithm Observatory, we are developing software that allows delivery workers to share trip and pay data directly with our department. This not only empowers workers and their organizations, but also gives DCWP an additional layer of visibility into a market the app companies would otherwise shield from scrutiny.
You will see the same approach in our response to consumer-facing harms. Consider customer service. We have already insisted on click-to-cancel subscriptions—developing the first municipal Click-to-Cancel Rule in America, which took effect October 1. We will bring that same proactive approach to AI customer service chatbots. In October, we will be announcing a proposed rule—the first such municipal rule in America—making clear that companies cannot trap consumers in endless chatbot doom loops when they are trying to cancel a service, dispute a charge, or obtain a refund. The system a company chooses to deploy must give people a meaningful way to exercise their rights.
The principle is straightforward: New technology does not create an exemption from old obligations. That philosophy will drive all of our enforcement work. But enforcement alone cannot answer every question technology is forcing us to confront. As companies use new tools to individualize prices, wages, and terms of service, we also have to decide which parts of economic life should remain common—visible, understandable, and governed by rules that apply to everyone.
Defending Our Shared Economy
We know that emerging technology can make changes in the economy feel inevitable, as though the rest of us are just passive bystanders along for the ride. We are not. We can decide, as citizens and as New Yorkers, which parts of economic life should be governed by mutually understood terms, and how much power companies should have to tailor those terms to each person’s vulnerabilities.
That is the second prong of our strategy: preserving a shared marketplace in which workers and consumers can compare their treatment, recognize common interests, and exercise collective power.
Consider surveillance pricing. We have already seen how these pricing practices would have likely doomed the 1902 meat boycott. And it’s encouraging to see that the Trump Federal Trade Commission—which previously mocked surveillance pricing as a concept Lina Khan made up—is now proposing an enforcement policy statement to require disclosure when companies engage in this practice.
But if we’ve learned anything from the last quarter century, it’s that a legal regime that places all the responsibility on consumers—the so-called notice-and-consent framework that pretended consumers could protect themselves by reading privacy policies and clicking “I agree”—was a profound failure of governance.
When I was at the FTC, we focused on prohibiting harmful practices, rather than letting companies disclose them away—we brought actions to ban the sale of sensitive health data, to ban biased facial recognition systems, and to destroy algorithms trained on ill-gotten data. We are taking the same approach in New York City.
That is why we have significant concerns with the FTC’s continued affinity for more disclosures, which hark back to the agency’s embrace of notice-and-consent 25 years ago. We believe surveillance pricing should be banned, not disclosed. And we support pairing that ban with a broad private right of action to ensure that companies are accountable for algorithmic manipulations.
We are also scrutinizing how agentic AI threatens our shared economy. Consumer protection enforcement often depends on identifying a practice that affects many people: a deceptive advertisement, a hidden fee, an abusive contract term. But when automated agents select products, negotiate terms, and make purchases on our behalf, each consumer could encounter a different set of representations, prices, and conditions—all through exchanges they may never see. The underlying misconduct may be systematic, while the evidence is scattered across thousands of private interactions. That makes it harder for consumers to recognize a shared grievance, and harder for enforcers to uncover the practice behind it.
That is why, later this fall, DCWP will be announcing a study on agentic AI led by professor Luke Herrine of Brooklyn Law School. Consumer protection law clearly applies when companies use these systems to deceive, exploit, or harm consumers. But agentic AI raises a broader challenge: It can make each consumer’s transaction increasingly individualized and increasingly difficult to compare with anyone else’s.
Our study will examine how to protect consumers in that environment—including what consumers should be able to expect from an agent acting on their behalf, how we should address conflicts of interest, what authority agents should have to commit consumers’ money or share their information, and what recourse consumers should have when something goes wrong. The goal is not only to protect consumers one by one but also to preserve their ability to recognize common practices, compare experiences, and act together.
Across the board, we are adapting our tools—and proposing new ones when necessary—to draw lines around the ability of powerful firms to make the essential terms of a transaction invisible, endlessly adjustable, and different for every person on the other side.
Investing in Collective Institutions
Our third task is to strengthen the institutions that can make those rights effective—and give working people a greater role in setting the terms themselves.
Our DoorDash case began with workers who courageously came forward about their pay. They were able to do that because they were not acting alone—they were organizing through the Workers Justice Project. Their experiences gave our investigators a place to start. And what began with a small group of organized workers ultimately became the largest labor enforcement action in city history.
This case shows what effective institutions can do together. Workers on the ground knew something felt off. Their advocates helped identify patterns across individual experiences. And government had the authority and capacity to compel answers and secure relief.
Under this administration, we are investing in expanding that capacity further. The Mamdani administration recently announced the largest investment in consumer and worker protection in New York City history, and we are using that investment to build our new Research and Analytics Division.
The premise is simple: Corporations should not enjoy a monopoly on the expertise needed to explain how markets work, or on the economic research that shapes public policy. Government needs capacity to test industry claims, understand what is happening in the marketplace, and produce rigorous research of its own.
Our work on delivery worker tips shows why that matters. App companies argued that higher worker pay would cause customers to tip less. DCWP’s analysis found something different: Changes the companies themselves made to their own interfaces had driven down tips by more than $550 million. We worked with the City Council to prohibit those design tricks, and workers have already gained more than $100 million in additional tips in 2026.
Our work against these tactics did more than raise earnings. It called out a fiction that pitted consumers against workers. It exposed companies’ own role in driving down workers’ tips. And it showed why government needs the capacity to test industry claims and challenge industry conduct.
But government cannot do this work alone. Workers need power of their own.
For many workers, that means unions, with the ability to bargain collectively over pay, scheduling, and the introduction of new technology into their jobs. But even workers who cannot currently access collective bargaining need ways to organize, press their demands, and shape the laws and rules that govern their work.
Popular
“swipe left below to view more authors”Swipe →
At DCWP, we are doing more than enforcing workers’ rights—we are standing with workers as they fight to expand them. With the Teamsters fighting for passage of the Delivery Protection Act. With security officers demanding a minimum wage. With delivery and rideshare workers challenging unjust deactivations. With the nurses, hotel workers, care workers, and baristas organizing for a better life. And with all New Yorkers who deserve a private right of action to assert their rights on their own.
Mayor Zohran Mamdani recognized the critical role of worker organizing when he created the Mayor’s Office of Worker Power, led by longtime organizer Tony Perlstein.
Across our administration, we believe firmly that as technology reshapes our economy, workers need to be front and center at the decision-making table. We need institutions that do more than remedy harm after the fact—they must give working people a voice in setting the terms from the start.
This all leads me back to Monroe Street on the Lower East Side. The legacy of the 1902 boycott rippled outward, across generations. The daughters of the women who led the strike would help build New York’s garment workers’ movement not a decade later. What began as a fight to feed their families became a much larger struggle over wages, working conditions, and who would have a voice in the city’s economic life.
That is the ambition we should bring to our work today. The cases we bring should recover what people are owed. The laws we pass should stop abuses before they spread. But our work must also leave something behind: a worker who knows her rights before her boss tests them, a tenant who knows how to organize her neighbors, institutions strong enough to turn individual grievances into collective action.
Every generation meets new forms of economic power. Ours faces technologies that can make that power more atomized, more opaque, and harder to challenge together.
We cannot predict every form artificial intelligence will take. Nor do we need to. What we need is to make sure that people, whatever the technology, can understand what is happening to them, compare their experiences with one another, and act together.
The women of 1902 left their daughters more than affordable food on the table. They left them organizations, habits of collective action, and an example of what working people could accomplish together.
In 2026, it’s time for us to rebuild those organizations, to relearn those habits of collective action, and to reimagine what working people can accomplish together.
More from The Nation
Trump’s refusal to heed the warnings from within the tech industry might end up being his most asinine legacy.
Sasha Abramsky
It has never been more urgent to revive the idea of peace, not as a slogan, but as the only serious, respectable, responsible, reasonable, humane position.
Katrina vanden Heuvel
While the president is unlikely to face any consequences for his patently illegal, taxpayer-funded ads, it’s essential to remember he should.
Elie Mystal
A call for freedom written across a pedestrian overpass, framed by wire.
OppArt
/
Jeanne Flavin
