18:12Machine-translated from Azerbaijani by AIRead in Azerbaijani
HUMXN to Pay for Home Repairs in Exchange for Filming to Train Robots
HUMXN is beginning to record home repair technicians at work to train AI-powered robots. The company will cover the cost of plumbing, electrical, heating, ventilation and air conditioning repairs. The program also includes free air duct and carpet cleaning. HUMXN is launching the program in Minneapolis, Chicago and Miami.
In exchange, the homeowner must consent to having the work recorded by a person wearing recording equipment. According to the company, it is difficult to accurately capture technicians’ natural movements in a laboratory. Footage recorded in homes will provide information about various types of equipment, faults, technicians and working conditions. The material will then be used to train AI robots.
The data collected will not be limited to the technician’s hand movements. AI models will learn how a fault is investigated, the order in which systems are checked and how the correct tool is selected. The technician’s response to changing conditions and how they verify whether the job has been completed successfully will also be recorded. The goal is to show robots every stage of physical work.
Homeowners who meet the program’s requirements will not pay for the service, while technicians will be paid for their work. In other words, HUMXN will cover the cost of the service in exchange for the right to film inside the home and collect data for training. The company selected these 3 cities specifically to collect data under different climate conditions. Residents can check whether they qualify for the program on a dedicated page.
HUMXN founder Thomas Unise believes that text and images available online are not sufficient to train AI robots. According to him, millions of skilled technicians perform complex tasks every day, but very little of this knowledge is preserved in a form from which AI systems can learn. The company wants to make that experience available to organizations developing intelligent machines. The technicians who generate the data are also expected to earn money from it.