In a workshop in Beijing, China, 40-year-old female worker Liu Yun wears a sensor bracelet on her wrist and small chips attached to her fingers as she repeats the motions of moving cups and wiping a dining table. Her daily job consists of repeating dishwashing and table wiping 200 times a day. The hand and arm movements collected in this way are utilized as training data for robots to learn household chores.

In China, robot trainers who teach human movements to artificial intelligence (AI) robots are emerging as a new job sector. As the competition to commercialize humanoid robots heats up, the value of behavioral data collected from real-life settings is increasing. However, a paradoxical phenomenon is also appearing where low-wage workers are hired on a large scale to develop robots that will eventually replace human labor.

Robot Trainer Recruitment Boom in China

According to the South China Morning Post (SCMP) on the 15th, advertisements recruiting robot trainers have been popping up recently on Chinese social media and job platforms. Monthly salaries generally range from 4,000 to 6,000 yuan (approximately 800,000 to 1,200,000 won). The positions are mainly applied for by the unemployed and young people seeking flexible work.

Liu Yun, who lost her job working in the automotive industry, struggled with re-employment due to her age of 40 before starting this work. She repeats the actions of moving cups and wiping surfaces according to instructions on a computer screen. In the workplace, workers' movements are monitored in real time, and productivity and accuracy rankings are displayed on a large screen.

Her daily work quota is 160 data samples. Skilled workers complete 260 or more. Liu Yun said that she felt her physical limits just from washing dishes and wiping tables 200 times a day.

A college student from Shanxi province worked as a robot trainer during summer vacation. He recorded standard movements for robots to follow by folding clothes and sweeping the floor. The daily wage was 120 yuan for the day shift and 135 yuan for the night shift. Due to the physical and mental burden of repetitive labor, many people quit the job after just a few weeks.

On the other hand, there are also evaluations that it is a new source of income that allows for telecommuting. Guo Hongjuan, living in Jiangsu province, works about six hours a day at home organizing books and recording item-moving movements to earn 4,000 yuan a month. The advantage is that she can work while taking care of her children.

Training data collection is expanding from homes to streets. Last August, Shenzhen-based technology company Pyson AI cooperated with an environmental cleaning company to have street cleaning workers wear data collection equipment. Even responses to electric bicycles passing by during cleaning are recorded so that robots can learn complex real-world environments.

The Paradox of the Labor Market: The Robot Trainer Hiring Boom

The reason Chinese companies are collecting human movements is that the core challenge of robot commercialization lies in securing real-world operational capabilities. Even if a robot possesses the skills to walk or dance, vast amounts of real-world data are required for it to pick up various objects and respond to unexpected situations.

According to Xinhua News Agency, in China, people are already directly operating robots in training grounds that replicate homes, factories, and logistics facilities to accumulate data.

Such data has also emerged as a new trade commodity. In January this year, Hubei province's Humanoid Innovation Center signed a training data sales contract worth thousands of hours with robot company Agibot. This means a specialized industry is forming not only for robot manufacturers, but also for data collection and processing.

Perspectives on the occupation of robot trainers are mixed. Self-deprecating expressions such as "cyber assembly line" or "telecommuting migrant labor" appear on Chinese online platforms regarding the fixation on repetitive tasks. There is concern that while new technologies create jobs, those jobs may remain stuck in high-intensity, low-wage labor.

One Chinese internet user pointed out, "Humans teach robots how to work, but someday those robots will replace the workers who taught them."

[Copyright (c) Global Economic Times. All Rights Reserved.]