David Tejeda helps deliver food and drinks to tables at a small restaurant in Dallas. And another in Sonoma County, California. Sometimes he lends a hand at a restaurant in Los Angeles too.
Tejeda does all this from his home in Belmont, California, by tracking the movements and vital signs of robots that roam around each establishment, bringing dishes from kitchen to table, and carrying back dirty dishes.
Sometimes he needs to help a lost robot reorient itself. “Sometimes it’s human error, someone moving the robot or something,” Tejeda says. “If I look through the camera and I say, ‘Oh, I see a wall that has a painting or certain landmarks,’ then I can localize it to face that landmark.”
Tejeda is part of a small but growing shadow workforce. Robots are taking on more kinds of blue-collar work, from driving forklifts and carrying freshly picked grapes to stocking shelves and waiting tables. Behind many of these robot systems are humans who help the machines perform difficult tasks or take over when they get confused. These people work from bedrooms, couches, and kitchen tables, a remote labor force that reaches into the physical world.
The need for humans to help the robots highlights the limits of artificial intelligence, and it suggests that people may still serve as a crucial cog in future automation.
“The more automation you inject into a scenario, the more, at least for now, you need those humans there to handle all the exceptions and just watch and supervise,” says Matt Beane, an assistant professor at the University of California, Santa Barbara, who studies robotic automation of manual work.
Human operators have been a feature of some commercial robotic systems for more than a decade. A few years ago, as new robots emerged in different workplaces, it seemed as if human helpers might be just a stopgap, helping until AI improves enough for robots to do things for themselves.
Now, Beane says, it seems that this workforce will continue to grow. “They’re cleaning up after the robot,” he says. “They are the human glue that allows that system to function at 99.96 percent reliability, according to reports given to some VP of automation somewhere.”
Beane says the smartest companies will use input from human operators to improve the AI algorithms that control their robots most of the time. Each time a person labels an object—a chair for example—in an image, it can help train the machine-learning algorithm that the robot uses to navigate.
But training AI this way is challenging, and there seems to be no shortage of new tasks for people to do. Beane says he has yet to come across a company that has successfully replaced human operators by having them train an AI algorithm.
Tejeda works for a company called Bear Robotics. The company’s cofounder and chief operating officer, Juan Higueros, says it is ramping up production of robots to meet growing demand, and also plans to hire dozens more robot operators.
“I do think this is going to become a very important aspect of how robotics companies that are in both structured and unstructured environments are going to have to operate,” Higueros says. He says the company has found an ample supply of workers in pockets of the US, including Texas and Utah.