What is it about robots and pizza?
A robot that bakes the well-loved meal was a big hit at the Consumer Electronics Show in Las Vegas last month, wowing onlookers as it prepared 300 pies an hour. The developer, a Seattle-based startup called Picnic, insists that pizza is only the beginning: “Our system will soon be able to make a wide variety of foods including sandwiches, salads, bowls, and more.”
And why not? After all, robot bartenders, including for the home, are all the rage (although the top-of-the-line model will set you back the price of a luxury car). Moreover, with labor costs rising and margins falling, no one doubts that the future of the food service industry is automation, both behind the counter and in the kitchen.
Still, there seems to be something special about robots and pizza. Domino’s, the world’s biggest pizza chain, touts its “autonomous delivery vehicle,” which is being tested in Houston. Big rival Little Caesar holds a patent on its own “apparatus … for assembling pizza” — that is, a pizza-making robot. And learned papers in serious journals explain the best way to teach an autonomous system to roll pizza … or even to select ingredients.
Well, of course. The world market for freshly baked pizza is nearing $150 billion a year, according to a 2019 report in PMQ Pizza Magazine, an industry trade publication. The largest share — over $50 billion — is in the U.S., but the fastest growth over the next five years is expected in Asia, Latin America and Eastern Europe. In China alone, annual sales are forecast to grow 21.6% by 2024.
That’s a lot of demand, but there’s also a lot of competition. For sellers of freshly baked pizza, as for others in the food service industry, rising labor costs and thinning margins make increased automation inevitable. Still, there’s no reason just yet for technophobes to panic — or, for that matter, for technophiles to celebrate. We have some time yet before being overwhelmed by the pizza-robot apocalypse.
Picnic’s automaton is undeniably fun and impressive to watch. It looks good too, sleek and unindustrial. One observer wrote that the robot resembles “a white, kitchen-sized iPhone.” It’s easy to see why the device has earned such fawning coverage — and why Picnic has already signed up some big customers.
Still, there are reasons to be cautious in our enthusiasm. First among them is the elephant in the room: the cautionary tale of Zume. Headquartered in Mountain View, California, Zume was a legitimate unicorn, with the clever idea of baking pizzas in special ovens inside its trucks, using GPS technology to determine when to heat the pies so that they would arrive fresh and hot at the customer’s doorstep. SoftBank’s single investment in the startup in 2018, which was greater than the total amount raised by all other robot-food-preparation startups combined over the previous five years, implied a value for the company of over $2 billion. The company even earned a cameo on the HBO drama “Silicon Valley.” In November, Vox Recode predicted that Zume might soon be worth $4 billion.
Last month, Zume shuttered its pizza operation, cutting 172 jobs in Mountain View and 80 more in San Francisco. Going forward, Zume plans to focus on sustainable food packaging.
Here’s a second reason for caution: Despite heady media claims that pizza is now being made by artificial intelligence (and a similar suggestion from the company itself), Picnic’s device doesn’t quite earn that encomium. I tend to agree with Melanie Mitchell of the Sante Fe Institute, who argues in her excellent 2019 book “Artificial Intelligence: A Guide for Thinking Humans” that what’s crucial to humanlike thought is the ability to apply common sense, a challenge researchers find difficult. We should avoid crying “AI” every time a device masters a particular task previously done by humans, even when the task is complex, like playing chess … or making pizza.
And Picnic’s robot, though quite ingenious, isn’t artificially intelligent just yet. Here’s a report from Seattle radio station KIRO:
“The human touch is still very much required for the machine to function. It has a vision system that reads the size and shape of the dough, and places the toppings and sauce on using a conveyor belt, but the dough and sauce must be made by actual human beings, who also have to put the pizza in the oven.”
This is impressive but not revolutionary. Automated assembly lines for automobiles do nearly as much, and have been around longer. As KIRO puts it, “The pizza-making robot is really good at making the easiest part of the pizza.”
But a system need not be artificially intelligent to be efficient and useful, and it’s easy to believe that Picnic is pointing the way toward the future. Picnic certainly isn’t Zume. For one thing, Picnic’s system is simpler. It lacks articulating arms and consists instead of “a series of modular, customizable food dispensers.” Even if the device currently performs only the easiest part of the operation, one assumes that its capabilities will improve over time. For another, unlike Zume, Picnic isn’t in the pizza business; it’s in the robot business.(Specifically, the company leases the device to customers, tailoring it to their needs and setting it up at no charge, all in return for a monthly fee.)
When the news from CES hit the web, Connecticut Sen. Chris Murphy took to Twitter to tell the world that he’s unimpressed: “It already feels like pizza made outside Connecticut/New Haven is made by mindless robots anyway, so I’m not sure how this would affect 99% of the country.”
Cute. But with so many restaurants struggling to pay their bills, and with technology improving all the time, it’s fair to say that the future of pizza (and of the rest of food service) belongs to the bots.
Stephen L. Carter is a Bloomberg Opinion columnist. He is a professor of law at Yale University and was a clerk to U.S. Supreme Court Justice Thurgood Marshall. His novels include “The Emperor of Ocean Park,” and his latest nonfiction book is “Invisible: The Forgotten Story of the Black Woman Lawyer Who Took Down America’s Most Powerful Mobster.”