Portuguese 
Spanish 
An AI startup has started testing a model that transforms household chores into data for robotics, with recordings made inside homes during free cleaning services in New York.
The artificial intelligence training startup Shift has begun offering free home cleanings in New York, USA, in exchange for permission to record the work done inside the homes.
The initiative, announced at the end of May 2026, involves using cameras attached to the professionals’ heads to record, in first person, tasks such as washing dishes, scrubbing bathrooms, mopping floors, organizing kitchens, and folding clothes.
According to the company, the videos captured during the services will be used to train AI systems and domestic robots capable of performing household tasks in the future.
-
Bangladeshi Student Transforms E-Waste into Solar-Powered Water Purifier for Flooded, Off-Grid Areas
-
Rare Diamond Found in Brazil’s Mato Grosso Formed 800 Kilometers Deep, Holds Clues to Earth’s Water Transport
-
Smart Connectivity Transforms Daily Life: High-Speed Internet, 5G Mobility, and Seamless Service Control
-
World’s First Fully Robot-Operated Hotel to Begin Trials with Real Guests in 2026, Full Opening in China Set for 2027
On the Shift website, the proposal is presented with the phrase “Your home. Cleaned for free.”, referring to the cleaning service without direct charge to the resident.
How Shift’s free cleaning works
Shift claims to connect New York residents with independent cleaning professionals.
Scheduling is done through the company’s platform, and the resident needs to be present at the start of the service to receive the professionals, explain what they want cleaned, and authorize the recording of the session.
The list of services disclosed by the company includes routine tasks such as trash removal, vacuuming, surface cleaning, refrigerator organization, cabinet tidying, sink cleaning, stove, bathroom, and shower cleaning.
The platform also mentions changing bed linens, pantry organization, mirror cleaning, and small tidying services.
The free service is linked to the use of data generated during the service.
According to Shift, the images recorded during the cleaning have commercial value for AI and robotics training, which allows offering the service without direct charge for a limited time.
In the terms published by the company, authorization for recording is a condition for the resident to receive the preferential price, which can be free or discounted, depending on the reservation.
Shift also informs that the value of each service is displayed before the confirmation of the scheduling.
According to the company’s website, professionals bring basic supplies, but the company recommends that the resident provide their own products if they have specific preferences.
Shift also requests payment data in some situations, such as when the resident is not present upon the arrival of professionals or refuses the service after the team’s travel.
Domestic data in AI training
Shift’s interest is in records of physical tasks performed in real environments.
Unlike text bases or static images used in AI model training, this type of recording shows movements, objects, surfaces, and sequences of actions within a house.
The company claims that more challenging cleaning environments can be useful for this type of training.
Even so, according to Shift itself, professionals can refuse specific tasks if they do not feel comfortable performing them.
In practice, simple activities for humans, such as washing a cup, folding a t-shirt, or cleaning a countertop, require automated systems to recognize objects, adjust movements, and deal with different environmental conditions.
This is the type of record the company says it seeks when recording cleaning sessions.
The model adds to an expanding market of data collection for AI.
International reports have shown companies using temporary professionals or paid collaborators to record everyday tasks, evaluate responses from automated systems, and produce training bases for artificial intelligence laboratories.
Privacy in homes recorded for AI
Recording inside homes requires the express authorization of the resident.
In Shift’s terms, the user declares that they have the right to allow recording in the property and assumes the responsibility of informing other adults present that the session will be recorded.
The company states that it anonymizes the images, processes the content, and licenses the material for AI and robotics training.
According to Shift, the content is not publicly disclosed or used for advertising.
The company also says it blurs names, faces, and other personal information before using the videos.
The protection informed by the company includes screens, identification documents, papers, cell phones, and other items that may reveal personal data during recording.
The platform’s terms indicate that consent can be withdrawn by the user.
However, Shift itself informs that the removal of a session may be limited after the material has been de-identified and incorporated into a database made available to third parties.
The company also states that the recording should be restricted to the work performed and the immediate surroundings of the activity.
According to the platform’s terms, professionals are advised not to record minors, intimate or medical situations, and sensitive identifiers.
Independent professionals and platform operation
Shift informs that the workers sent to residences are independent professionals evaluated by partners.
The company states that they are not direct employees of the platform.
In the terms of use, the company presents itself as a technology platform that facilitates scheduling and communication between the client and the provider.
Shift states that it does not directly perform the cleaning service, does not act as an employer, and does not function as a staffing agency.
This distinction is relevant because it defines how the company describes its participation in the operation.
The service contract, according to the terms, is formed between the client and the provider identified at the time of booking.
According to a Business Insider report, Shift is the consumer brand of microagi, an AI research lab based in Germany.
The publication reported that microagi was founded by Bercan Kilic, Yoan Iliev, and Anton Poletaev.
The same report cited Harry Kilberg, general manager of Shift in the United States, stating that the company operates in 15 countries and has 14,000 operators collecting real-world data.
The statement was attributed to Kilberg by Business Insider.
Demand for the service in New York
The proposal gained attention in the United States after the campaign was publicized on social networks and in technology-specialized media.
According to Business Insider, the first 250 sessions were quickly booked, and the company recorded thousands of scheduling attempts.
Shift states that the free cleaning is available for a limited time.
The service initially appears concentrated in New York, while international reports point to expansion plans for other cities.
The Verge reported that the company cited San Francisco, London, Zurich, and Munich as possible next markets.
The report also stated that the company intends to apply the model to other physical activities, such as cooking, plumbing, and construction.
The advancement of this type of initiative occurs at a time when AI companies are seeking data from tasks performed outside digital environments.
Instead of just labeling texts or images, part of the sector is trying to record how humans perform actions in the physical world.
For residents, the offer presents an objective condition: receiving a cleaning without direct payment in exchange for recording the service inside the home.
For AI and robotics companies, the videos can form training bases aimed at systems that attempt to reproduce household tasks.
Shift’s proposal, therefore, combines an everyday service with a form of data collection for technology.
The decision to participate depends on the resident’s authorization and the rules provided by the company regarding recording, anonymization, and use of the material.
