David Tejeda helps delivering food and drink to the tables of a small Dallas restaurant. And another in Sonoma County, California. Sometimes he also helps out in a restaurant in Los Angeles.
Tejeda does all of this from her home in Belmont, Calif., Tracking the movements and vital signs of the robots that roam each establishment, bringing dishes from the kitchen to the table and bringing back the 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 say, ‘Oh, I see a wall that has a painting or some landmarks,’ then I can locate it to face that landmark.”
Tejeda is part of a small but growing ghost workforce. Robots accept more types of blue collar work, forklift driving and carrying freshly picked grapes at storage shelves and waiting tables. Behind many of these robotic systems are humans who help machines perform difficult tasks or take over when confused. These people work from bedrooms, sofas and kitchen tables, a distant workforce entering the physical world.
The need for humans to help robots highlights the limits of artificial intelligence, and this suggests that people can still serve as a crucial cog in future automation.
“The more automation you inject into a scenario, the more you need, at least for now, these humans 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 labor.
Human operators have been a feature of some commercial robotic systems for over a decade. A few years ago, as new robots appeared in different workplaces, it seemed like human assistants could be just a stopgap, helping until the AI improves enough for the robots to do things on their own.
Now, says Beane, it looks like that workforce will continue to grow. “They clean up after the robot,” he says. “They are the human glue that keeps this system running with 99.96% reliability, according to reports given to a VP of automation somewhere.”
Beane Says Smarter Businesses Will Use Contributions From Human Operators To Improve AI algorithms who control their robots most of the time. Whenever a person names an object (such as a chair) in a picture, it can help shape the machine learning algorithm that the robot uses to navigate.
But training AI in this way is a challenge, and there is no shortage of new tasks. Beane says he has yet to meet a company that has successfully replaced human operators by asking them to train an AI algorithm.
Tejeda works for a company called Robotic bear. The company’s co-founder and chief operating officer Juan Higueros said he was ramping up robot production to meet growing demand and that he also plans to hire dozens of other robot operators.
“I 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 a sufficient supply of workers in pockets across the United States, including Texas and Utah.
Remote robot work is a growing category in job postings, especially in robotics startups looking to place systems in new contexts that present challenges for AI. Perceiving, interpreting and operating in an ever-changing environment remains an unresolved problem in AI and robotics, despite impressive progress in recent years.
Interest from some problem-focused startups is another sign that remote robot fights are taking off. Jeff Linnell, who previously worked on robotics at Google, left to found Forming in 2017, when he realized that a more remote operation would be necessary. “There are all kinds of applications where a robot can do 95% of the mission and one person can take over,” he says. “This is our thesis.”
Formant’s software combines tools to manage robot fleets with others to set up remote robot operator teams. “The only way to achieve an economy of scale over the next decade, in my opinion, is to have a human behind him, running a fleet,” he says.