The data you share has a way of coming back to bite you.
In the apartments of Manhattan, a quiet transaction is unfolding: residents receive free cleaning services while cameras record every human gesture, feeding a machine learning system designed to render those same gestures unnecessary. Micro AGI's program, called Shift, sits at the intersection of convenience and surveillance, offering something tangible in exchange for something whose true value remains difficult to measure. It is a story as old as labor itself — who benefits from work, who owns its knowledge, and what is lost when the hand that learns is no longer human.
- A New York startup is offering free apartment cleaning in exchange for intimate video footage of human hands at work, captured by camera-equipped cleaners and fed directly into robot-training systems.
- Privacy advocates warn that what feels like a fair trade conceals a profound imbalance — the data collected inside your home is worth far more than a clean kitchen, and its future uses remain beyond your control.
- The company defends itself by pointing to a landscape already saturated with covert data extraction, arguing that at least Shift makes the exchange explicit and gives participants something real in return.
- Beneath the debate over consent and transparency lies a harder question: the workers generating this footage may be training the very machines that will one day replace them, and few seem troubled by the irony.
On a Tuesday morning on Manhattan's Upper East Side, two young men arrived at an apartment carrying cleaning supplies and cameras mounted to their caps. They were employees of Micro AGI, and their presence was the product of a program called Shift — a service that offers New Yorkers free apartment cleaning in exchange for something less visible: footage of every movement the cleaners make, recorded and used to train autonomous robots.
The logic is straightforward. Founder Bercan Kilic believes that if machines are to navigate the unpredictability of real homes — varying layouts, shifting light, different tools — they need vast quantities of real-world data. Shift collects that footage, anonymizes it, and sells it to robotics and AI companies. The cleaners, mostly college graduates, work five apartments a day and are paid above market rate. One even sent a recording kit to his mother so she could contribute her own domestic movements to the dataset.
Privacy experts are not reassured. The Electronic Frontier Foundation's Rory Mir warns that data shared today has a way of resurfacing tomorrow — shared with other businesses, sold to governments, or used in ways no one anticipated. Calli Schroeder of the Electronic Privacy Information Center went further, calling Shift's model a creative form of privacy invasion, noting that in-home recordings capture far more than most people realize: routines, possessions, the intimate geography of private life. A free cleaning, she argues, is a poor price for what is being compiled and sold.
Kilic counters that Shift is more transparent than the social media platforms already harvesting data without offering anything in return. The transaction is visible, he says, and participation is voluntary. What remains unresolved is whether the exchange is truly fair — and what becomes of the workers whose labor is teaching machines to make them obsolete.
On a Tuesday morning in Manhattan's Upper East Side, two young men in their mid-twenties arrived at an apartment with cleaning supplies and something unusual: cameras mounted to their caps, wired to their phones. They were not filming a documentary. They were not part of a startup stunt. They were employees of Micro AGI, a company that has figured out how to get New Yorkers to let robots watch them live—by offering to clean their homes for free.
The program is called Shift. The pitch is straightforward: you get a spotless apartment at no cost. In exchange, every movement the cleaners make—every swipe of a cloth, every adjustment of a tool, every interaction with the objects in your space—gets recorded and fed into a machine learning system designed to train the next generation of autonomous robots. The company is betting that if it can collect enough footage of human hands doing real work in real homes, it can teach machines to do the same thing. Shift's founder, Bercan Kilic, told the BBC that the goal is nothing less than advancing humanity. He acknowledged that the challenge is immense: kitchens differ, lighting changes, tools vary, and nothing stays the same for long. To build robots that can adapt to this chaos, the company needs what he called "tonnes" of data.
The business model is elegant and unsettling in equal measure. Micro AGI collects the footage, anonymizes it, and sells it to robotics companies and other AI firms hungry for training material. The cleaners themselves—mostly college graduates cycling through the startup world—work five apartments a day, five days a week, stationed in New York indefinitely because demand for free cleaning is high. They are paid above the going rate for cleaners in the city, and some of them seem genuinely excited about being part of the AI revolution. One cleaner even sent a recording kit home to his mother, who now films herself doing household tasks from her own perspective, contributing to the dataset.
But privacy advocates are alarmed. Rory Mir, director of open access and tech community engagement at the Electronic Frontier Foundation, sees Shift as part of a troubling pattern. Companies are increasingly offering what looks like a transaction—a free service in exchange for data—but the math is deceptive. "The data you share has a way of coming back to bite you," Mir said. Even if you trust the company collecting it today, there is always the risk that the information gets shared with other businesses or governments tomorrow. We have spent decades watching our data used to manipulate us through advertising and predatory pricing schemes. This feels like the next chapter.
Calli Schroeder, director of the AI and human rights program at the Electronic Privacy Information Center, called Shift's approach "a diabolically creative way to sell privacy invasion." She pointed out that in-home recordings capture far more sensitive information than most people realize—the layout of your apartment, your daily routines, your possessions, the way you move through private space. And while a free cleaning might feel like fair compensation, it is, in her view, a pittance compared to the value of the data being compiled and sold. The technology being developed from this footage could eventually displace the very workers who are generating it.
Kilic pushes back on the criticism. He argues that Shift is actually more honest than the platforms most people use every day. Your data is being harvested constantly by social media companies and websites, he points out, and you are not being paid anything for it. At least with Shift, there is a clear transaction: you get something tangible in return. If you do not want to participate, you do not have to. "We don't expect everyone to like it," he said, "and that is fine."
What remains unresolved is whether the trade-off is actually fair, and what happens when the robots are ready. The cleaners working for Shift today seem unbothered by the prospect of their own obsolescence. They speak about embracing AI early as a hedge against the future. But the broader question—whether a free apartment cleaning justifies handing over footage of your home to a company that will profit from it—is one each New Yorker answering the door will have to answer for themselves.
Citas Notables
Models need to learn how their hands, cameras and environments work together in the real world, where every object and lighting condition is different.— Bercan Kilic, Shift founder
Even if you trust the business collecting it, there is always a risk of them sharing that information with other businesses or governments.— Rory Mir, Electronic Frontier Foundation
La Conversación del Hearth Otra perspectiva de la historia
Why would anyone agree to this? A free cleaning seems like a small thing compared to what you're giving up.
That's the design of it. The cleaning is real and immediate. The data sale is abstract and distant. Most people don't think about what a video of their hands, their apartment, their routines is actually worth.
But the company is being transparent about it, right? They're telling people what's happening.
They are, which is more than most tech companies do. But transparency and fairness are not the same thing. You can be honest about exploiting someone and still be exploiting them.
What's the real concern here? Is it that the data gets misused, or that people are undervaluing what they're trading away?
Both. But also this: once the robots work, the cleaners are gone. The data lives forever. The company profits indefinitely. The person who let them film their apartment gets a clean living room and nothing else.
So it's a one-time payment for perpetual value extraction.
Exactly. And the person making the trade probably doesn't think about it that way.