Startup ‘Shift’ Offers Free Cleaning in Exchange for Robot Training Data A new startup is turning housework into high-tech training data, offering free home cleaning while it films every scrub and sweep to teach future robots how to work inside people’s homes.

May 28–29: Shift’s unusual offer goes public

Late May, German tech firm MicroAGI began promoting its new Shift app, promising “free, trusted professional house cleaners” to New Yorkers in exchange for “first-person cleaning footage to help train the next generation of household robots.” Cleaners wear camera-equipped headgear—described by co-founder Bercan Kilic as a “magic hat” that records from their point of view—as they vacuum, mop, and wash dishes.

The company argues the value of this training data is high enough to subsidize the cleanings: “You get a spotless apartment. We get training data. Everyone wins.” The service launched first in New York, with plans to expand “very soon” to San Francisco, London, Zurich, and Munich.

How the data is used—and promised safeguards

Shift’s pitch sits in a broader scramble among tech companies for real-world footage of chores, as robots struggle with the messy physics of everyday life, unlike text-based AI that can be trained on scraped online content. Domestic videos of “scrubbing dishes, wiping counters, dusting tables, mopping floors” are seen as critical to overcoming that bottleneck.

Shift says privacy is “fully protected,” claiming “names, faces or other personal information is automatically anonymized, with any sensitive details blurred before it’s ever used,” using “advanced machine learning models” on capture devices to perform “irreversible transformations” before upload.

Growing benefits—and unease

Supporters frame the deal as a transparent trade: free services in return for explicitly requested data, part of a wider trend where companies “pay people for the real-world data needed to train their robots” rather than scraping it without consent.

Critics, though, question whether homes can ever truly be anonymized and note gaps, such as no clear way to have footage removed from training datasets once collected. As more firms adopt similar models, the core tension remains: are residents savvy participants in a data marketplace—or unwitting extras on a permanent training set for machines?

Continue reading https://foxvector.com/stories/019e76a8-3d66-1a97-7061-09e46650df99