Service Robots Roll Forward

History is filled with examples of robotic devices designed to reduce, eliminate, or improve upon human labor. From washing machines to roaming vacuum cleaners, various machines have transformed the way we work and live.

But now, far more sophisticated service robots are wheeling into the picture. These devices aim to take humans out of the labor loop and, in the process, improve the speed and efficiency of interactions. They can carry plates between a restaurant kitchen and diners’ tables, deliver a toothbrush to someone on the 28^th floor of a hotel, and ensure that a hospital patient receives her medications on schedule. They also are adept at stocking shelves, taking orders at a fast food restaurant and serving as emotional companions.

“The field of robotics is rapidly transitioning from classic ‘programmed automation’ to modern ‘AI-powered autonomy’,” says Christine Boles, vice president in the Network and Edge Group and general manager of Federal and Industrial Solution at Intel Corporation, which designs various hardware and software components used for robot systems. “Artificial intelligence (AI), sensors, vision, location, and wireless technologies are fueling this transition.”

Service robots, which are no stranger to Japan and other parts of Asia, where they already are employed as hotel receptionists, housekeeping attendants, window washers and more¹, are taking on new roles. There also are robots stocking shelves at convenience stores² and companion robots to keep people company.³ Now these devices are moving into the mainstream in restaurants, hotels, hospitals, and homes in the US and Europe.

Yet, there also are growing concerns about these machines eliminating jobs, reducing social interactions, misbehaving and collecting highly personal data about human health and habits. While there is growing consensus that service robots are here to stay—and they will change the workplace and our homes in significant ways—plenty of questions remain about what their exact role will be.

“The technology is advancing rapidly. Service robots may soon become part of everyday life,” says Zackory Erickson, an assistant professor in the Robotics Institute at Carnegie Mellon University.

At Your Service

Service robots—typically defined as autonomous and adaptable systems that interact, communicate, and handle various tasks for humans—have been around for more than two decades. Sony’s robotic dog Aibo, for instance, appeared in the late 1990s and the first Roomba’s became sweeping carpets in 2002.

But these devices are advancing rapidly. In the US, Hilton Hotels, Marriott Corporation, Denny’s, and Buffalo Wild Wings are just a few of the companies experimenting with robots—and integrating them into regular operations.⁴ A group of researchers from Chongqing University in China reported in 2022 that robot applications in hotels, tourism, catering and other service areas are “gradually emerging.” Areas such as express delivery, housekeeping, and security are prime candidates for the technology.⁵

One company at the forefront of the trend is Bear Robotics, a Redwood City, CA, manufacturer of systems used in restaurants, hotels, and hospitals. It has already sold more than 10,000 service robots. Its Servi+ model carries food to tables in restaurants and then hauls dirty dishes to the dishwasher. “A robot can deliver food to a table as soon as it’s ready and eliminate the need for staff to carry as much as 30 or 40 pounds of dishes back to the kitchen,” says co-founder and chief operating officer Juan Higueros. The Servi+ robot costs end-users about $1,000 per month, which equates to a wage of about $2.75 per hour.⁶

By reducing mundane and physically taxing work like carrying dishes to and from the kitchen, wait staff have more time to engage with customers. This translates into a qualitatively better experience—and possibly higher tips, Higueros argues. Bear’s robots function within an interconnected software platform that coordinates and orchestrates their activities. They incorporate cameras, LDAR, and other sensors that allow them to avoid running into objects, people, and other obstructions.⁷ “The algorithm maps an area and detects humans and objects—people, wheelchairs, diaper bags, whatever—and navigates around everything,” he explains.

Bear Robotics is not alone. United Robotics Group of Germany offers a 23-inch tall mobile humanoid robot named NAO⁶ that answers questions and interacts with people. The $13,000 system incorporates directional microphones, speech recognition, cameras, touch sensors, and 25 degrees of freedom.^8,9 United Robotics also sells a 4-foot-tall, 60-pound humanoid-like robot named Pepper that is designed to work as either a receptionist or as a mobile tour guide at hotels, banks, medical offices, or museums. It costs between about $32,000 and $38,000.¹⁰

Meanwhile, Austin, TX-based Diligent Robotics, has developed a mobile robot that supports patient care teams in medical facilities. Its Moxi robot delivers supplies, lab samples, medications and more within a facility. Moxi incorporates a robotic arm that allows it to operate elevators and doors, offers LED eye expressions that make it appear more human, and uses AI to help it quickly learn the layout of a hospital.¹¹ Still another autonomous service robot named Akjara, developed by an Irish startup, cleans and sanitizes patient rooms.¹²

“It isn’t so much a question of whether we can develop the technology to build sophisticated service robots,” Erickson says. “It’s whether we can design them in acceptable ways—and use them in a wide array of situations and environments that are less structured and sometimes random. If companies develop affordable robots that solve common tasks, we are going to see them rapidly become part of everyday life.”

Making It All Compute

Designing robots that can think, act, and operate like humans is no simple task, however. Service robots require “unique design and engineering considerations,” Intel’s Boles says. Part of the challenge is rooted in the autonomous nature of these machines. Aside from the need for built-in navigation and programming, service robots must frequently offload compute tasks to the cloud or edge—without incurring a performance hit. At the same time, there is a need for them to interact with lay people.

Building a complete control loop is all about a sense-act-think cycle, Boles says. This includes situational awareness and sensing powered by CPUs, GPUs and VPUs. Yet, it also is essential to focus on less obvious factors, such as mechanical noise level, charging breaks, and the need for specialized materials, such as germ-resistant surfaces for restaurants or home kitchens. What’s more, “Shape, size, and even the color of the robot must match the environment and the roles of people that interact with it,” she says.

To be sure, the factors that go into designing a highly functional service robot extend beyond technology. “One of the biggest obstacles is engineering service robots to work in existing spaces—restaurants, hospitals and hotels—that weren’t designed for robots,” Erickson says. There are also practical considerations, such as when, how, and why to use the robot. “If you work in a hotel, you probably don’t want to bother delivering a toothbrush that doesn’t lead to a significant tip. But you do want to deliver the $100 bottle of champagne.”

The goal, Higueros says, is to assist and complement humans, rather than replace them. Of course, reality most likely lies somewhere between aiding workers and replacing them with machines. A 2023 research study found that while human-robot collaboration can create new jobs and increase productivity, it can also lead to increased job insecurity, burnout, and workplace incivility.¹³ “Job displacement will occur. But, as work becomes more complex, more jobs typically appear,” says Joshua Conrad Jackson, an assistant professor in the Booth School of Business at the University of Chicago.

A bigger question revolves around the role robots should play in the everyday lives of humans. This includes concerns about whether a robot can be fully trusted to handle a task and whether it could cause some type of harm, Jackson says. Research shows that people are especially afraid of robots making moral decisions. As the stakes increase—from delivering packages in a modern high-rise apartment to serving as companions in rehab facilities or performing life-or-death surgical operations, concerns pile up. “People aren’t comfortable with robots in charge of things. The higher the stakes, the greater the level of fear,” Jackson says.

High levels of mistrust have led researchers and engineers down the path of anthropomorphism. Indeed, getting human-robot interactions right requires more than state-of-the-art guidance technology and AI. “It’s all about understanding how to make these machines more relatable,” Jackson explains. “Do you design a robot with a nervous laugh? Do you imbue it with human qualities like embarrassment or shame? Do you program it to make intentional errors and then apologize? If so, what errors are appropriate? The historical goal of computers and robotics is to avoid errors,” he points out.

The specific characteristics of a robot also matter. For example, female names, voices, and other expectations associated with empathy resonate with users more than male characteristics. In addition, a wider range of emotions is generally viewed more favorably and when the machines display feelings, they are more readily accepted.¹⁴ Jackson suspects that researchers eventually will learn how to perfect robot behavior—and even make the devices adaptable to the peculiarities of different cultures and groups. “The Uncanny Valley effect may become irrelevant,” he says.

Values Matter

Just as robotic devices have changed the face of factories, warehouses, and even surgical units, service robots will alter business models and consumer behavior in the years ahead, says Carolina Salge, an assistant professor of Management Information Systems at the University of Georgia’s Terry College of Business. “They are here to stay, and they will span many industries and functions. We’re just at the beginning of learning about their potential impact,” she says.

Service robots introduce an array of ethical questions and concerns. These include their role in society, potential hacking (many lack adequate security controls) and deliberate program modifications that could lead to failures. Not surprisingly, these concerns escalate where robotic systems handle highly sensitive data, such as in homes and hospitals.

In addition, there are concerns about behavioral biases that service robots can pick up during AI training, along with how these machines alter interactions among people. “One of the things that tends to get overlooked is how robotic devices—as they evolve from Roombas to sophisticated companions—have the potential to decrease human contact and social interaction. This could lead to negative outcomes,” Salge says.

Some businesses have discovered that service robots introduce more problems than they solve. These highly mechanical systems sometimes break down or underperform, and managing and servicing a fleet of robotics can lead to higher costs than employing humans. In Japan, for instance, the Henn-na Hotel laid off half of its robot staff in 2019 due to an array of practical problems and high costs associated with repairs.¹⁵

Nevertheless, service robots are continuing to roll forward at a rapid pace. More advanced electromechanical components, better sensors and actuators, persistent connectivity, longer-lasting batteries, and more advanced algorithms, along with advances in large language models and generative AI, are delivering the technical underpinnings for more advanced and even human-like systems. At the same time, engineers and designers are learning how to build spaces with commentary devices and systems that allow robots to work next to people.

Fears over robots replacing humans are greatly exaggerated, argues Gerald C. Kane, professor and the C. Herman and Mary Virginia Terry Chair in Business Administration at the Terry College of Business at the University of Georgia. “Humans have a long history of turning to new forms of media—books, television, social media and more—as surrogates for human relationships. In all likelihood, physical robots and online bots will interact with humans in new ways—and lead to new ways to interact. They will become more intertwined with human work…and our daily lives.”