AI Is Taking Cashier Jobs—And People Are Helping Remotely

Robots in Stores? That’s Just Daily Life in Tokyo

Workforce specialists have long discussed how remote work completely transformed certain industries. This is particularly evident in IT, where an American company might remotely hire a Pole, an Indian, and a Spaniard, all of whom work seamlessly on one project from their home countries.

For years, physical laborers claimed this trend wouldn’t affect them. Warehouse workers, store cashiers, and production line staff—you can’t do those jobs remotely, right? However, what’s happening now in the Japanese labor market proves otherwise. To understand this, we must travel 3,000 kilometers south of crowded Tokyo, to Manila, the capital of the Philippines.

Behind the Counter Sits… Someone in Manila

In a multi-story office building in Manila’s financial district, about 60 young men and women monitor AI-powered robots that replenish shelves in stores in distant Japan. When a bot drops a can of soda, the operator puts on Virtual Reality (VR) goggles and uses controllers to help the robot pick it up.

Tokyo-based company Telexistence is responsible for the entire system. They built the robots and placed them in over 300 stores across Tokyo starting in 2022. How can remotely controlling robotic cashiers possibly pay off? Wouldn’t it be better to hire someone locally?

AI at the Checkout Is Faster, Cheaper, and Never Stops

“It’s hard to find shelf-stocking workers in Japan. And if you do find someone willing, it will be very expensive. The minimum wage there is high,” says Juan Paolo Villonco, founder of Astro Robotics, the company employing the operators in Manila, in an article quoted by Rest of the World.

In Japan, the average minimum wage (with regional variations) converts to nearly $2,400 USD per month. Meanwhile, Astro Robotics employees earn a maximum equivalent of only $300 USD monthly. The calculation is simple, especially since a Japanese cashier can only be in one store at a time, while one Filipino operator supervises up to 50 robots simultaneously.

Robots Learn from Humans, Only to Replace Them

The machines operate completely autonomously for 96 percent of the time. Operator intervention is only required when they make a mistake. During an eight-hour shift, this happens about 50 times, with each intervention lasting up to 5 minutes.

Rowel Atienza, a machine learning professor at the University of the Philippines, highlights the negative effects of repeatedly putting on VR goggles. “This can be truly exhausting. Imagine teleportation, a sudden disconnection from your surroundings, a change in level—all of this can cause accidents,” he argues.

The Human as the Robot’s Shadow

Lionel Robert, a Robotics Professor at the University of Michigan, draws attention to another dimension: “Previously, people lost their jobs to a machine. Now, they have become the observers of the machine doing that job. You are like the robot’s deputy.” Indeed, this reality could be more damaging to self-esteem than the mere loss of a job.

They Are Training Their Own Successors

The Filipino employees of Astro Robotics are undermining not only the Japanese labor market but also their own employment. Data from their interventions is collected by Telexistence. In June 2025, the company announced it was transferring this information to Physical Intelligence, an AI firm operating in San Francisco. Why?

AI Devours Its Own Tail

The American company will use the data to better train the programs that control the robots. Telexistence directly stated in its press release that this collaboration “aims to transform these manual remote control tasks into fully autonomous operations.”

The Future of remote robot jobs? Steering Robots from Across the Globe

Already, the “pilots” of the cashiers in the Philippines spend only a small part of their shift manually correcting the robots’ errors. However, the data from these interventions may soon mean these devices won’t require human supervision at all. The Manila operators are training their successors every day—successors who are more convenient, trouble-free, and cheaper to employ. It is only a matter of time before robots replace the humans currently supervising their remote robot jobs.

Read this article in Polish: AI zabiera pracę kasjerom. A ludzie pomagają… zdalnie