Skip to Content
Artificial intelligence

Robots Are Invading Malls (and Sidewalks) Near You

A new generation of robots is heading out of the factory and into urban settings to help you get packages and snacks.

At the upscale Stanford Shopping Center in Palo Alto, California, people are taking selfies with a roving robot that looks like a cross between Wall-E’s girlfriend and R2D2. It’s actually a K5 robot security guard—a 300-pound, sensor-filled droid made by a startup called Knightscope that patrols the area and detects suspicious behavior.

K5 is part of a small but growing number of human-scale mobile robots that are finding employment outside the confines of industrial settings like factories. They’re invading consumer spaces including retail stores, hotels, and sidewalks in a quest to deliver services alongside human staff members for a fraction of the price of employing people to do a variety of typically unexciting tasks. The machines come with navigation capabilities and safety features to allow them to perform simple jobs autonomously without putting people at risk.

“This is not about replacing people, but augmenting them,” says Steve Cousins, CEO of Savioke, which makes a room-service robot called Relay, also known as Botlr.

Relay robots currently operate in eight U.S. hotels, making around 25 to 30 daily trips to guest rooms carrying requested items including snacks, toothpaste, and packages.

The K5, made by Knightscope, is a robotic security guard.

At the Aloft hotel in Cupertino, California, the 36-inch-tall Botlr docks behind the reception desk until a guest calls for an item. A staff member then places the object—a bag of chips, for instance—inside the robot’s 21-liter drum, closes the lid, and uses its touch screen to key in the room number. Then it whizzes off to the elevator—which has been modified so the robot can summon it wirelessly—before navigating to the correct room by using lidar and depth-sensing technology. The Botlr waits until the guest opens the door before opening its lid to reveal the drum’s contents, and then heads back to its charging station downstairs (if the guest doesn’t answer the door, as apparently happened when I visited the hotel, the robot simply brings the item back to its dock).

To test the Botlr’s mettle, Cousins sent it on a mock mission within the hotel’s lobby while we purposefully got in its way; it navigated around us very well. 

Tally, a robot made by Simbe Robotics, can patrol store aisles taking inventory.

Robots have been mingling with humans in several stores, too, including a Target in San Francisco, where a robot called Tally was used for a trial in which it trundled up and down aisles carrying out inventory checks—a mind-numbing task for humans. Tally detects when products are out of stock or moved so staff know to replace them. According to Tally’s creator, a startup called Simbe Robotics, it can complete an audit of a medium-sized store in around half an hour, with 96 percent accuracy. The same task would take a human 25 hours, and the company contends people are only about 65 percent accurate.

Simbe Robotics CEO Brad Bogolea says that in order to make shoppers feel comfortable with robots wandering around the store, it’s important that the robots don’t look threatening.

“We’re not making something that looks like the Terminator,” he says. “Most often consumers ask us if it’s building a map, cleaning the floor, or if it’s a security robot.”

Starship’s robots autonomously drive down sidewalks to deliver packages.

It’s a design ethos echoed by Starship, one of a number of startups trying to crack so-called last-mile delivery, which typically refers to a package’s journey from a delivery hub to your home or business. Starship’s six-wheeled self-driving robots have a cartoony aesthetic reminiscent of Google’s latest self-driving car model. Each uses nine cameras to navigate sidewalks, carrying packages or groceries inside a locked chamber that’s opened using a code sent to the recipient via SMS.

So far, 15 Starship robots have driven 3,200 miles on public sidewalks in Arkansas, London, Estonia, and the San Francisco Bay Area. I got to see one of them do some laps around a park in San Francisco, where it was funny to watch people do double takes as they tried to work out what it was and to whom it belonged.

Savioke’s Relay robot—also known as a Botlr—brings small items to hotel guests’ rooms.

Operating outdoors in urban settings is particularly challenging, says Allan Martinson, Starship’s chief operating officer. That’s because the company’s robots must drive around in open areas with unpredictable terrain, crowds of people, and lighting conditions. To help with any issues the robots may encounter—like traversing a crosswalk—they can be controlled by a human operator at any point.

Whenever robots interact with untrained humans, safety is a big concern. To avoid problems, these robots tend to have obstacle-detection technologies and low speed limits; Savioke and Starship, for instance, limit their robots to moving at four miles per hour.

Yet while all of these robots can safely interact with humans on a simple level, there’s plenty of room for improvement, says Carnegie Mellon computer science professor Manuela Veloso, who studies artificial intelligence and robotics. For instance, she thinks we should be able to teach them through instructions and corrections.

“Inevitably people will see robots that move too close to a wall or don’t say thank you and they’ll want to correct the behavior. The robots need to learn from these interactions,” she says.

Deep Dive

Artificial intelligence

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

Providing the right products at the right time with machine learning

Amid shifting customer needs, CPG enterprises look to machine learning to bolster their data strategy, says global head of MLOps and platforms at Kraft Heinz Company, Jorge Balestra.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.