The burger robot is cool as hell. The elegant mass of machinery, color-coordinated in white, copper, and blond wood, represents some important, unanswered questions, such as how automation will reshape the dining industry. But before we address that, let’s be real about San Francisco’s Creator restaurant and its burger-making machine: This thing is far out!
Unlike the automated woks at Spyce in Boston, which fry up pre-sliced and portioned ingredients, the machine at Creator is fed whole brisket and chuck, peeled onions, pickles, etc. It slices tomatoes; it bisects, toasts, and butters buns; grinds and sears beef; and squeezes out doses of ketchup measured to the milliliter.
Never mind that this enables Creator to use chef Heston Blumenthal’s ridiculously perfectionist methodology for a “loosely packed” and “vertically aligned” bite, meaning that the grain of the burger will run in the same direction as your teeth. If only for the execution of grinding beef to order, it is an impressive feat of innovation, and a sign of things to come.
The savings in human labor, according to the restaurant’s messaging, allows the restaurant to reallocate spending on higher quality, more sustainable ingredients, and offer better working conditions for staff. Though at present, Creator has over 60 employees and counting.
This marvel of the modern age cooks the most popular food item in America. So will the country’s biggest purveyor of hamburgers, McDonald’s, buy them out or attempt to replicate the technology?
And how long will it be until our burgers, burritos, and grain bowls are produced by machines?
While Creator founder Alex Vardakostas seems intent on building his company, rather than selling his technology, the big players have to be sniffing around. In 2019, 18 states increased the minimum wage, with many planning to raise it to $15 by 2025. So fast-food chains, for whom rising wages are an extinction-level meteor hurtling toward their profitability, are likely considering replacing workers with machines.
But robots will not replace cooks in all restaurants. The dining industry is tiered — fancy, luxury destinations for the wealthy (and regular folk on special occasions); mid-range, full-service restaurants; and fast-food, aka QSR (quick service restaurants), marked by counter ordering versus table service. At the high end, don’t expect the Terminator to be recommending wine pairings anytime soon. In the mid-range, R2-D2 isn’t going to be inspired to concoct a lunch special from the zucchini blossoms growing in the restaurant’s rooftop garden. However, the fast-food level, where profitability has always been about sourcing the cheapest ingredients or finding efficiencies in assembly-line cooking, is ripe for automation that will displace jobs.
Vardakostas insisted he is not looking to sell to the highest bidder.
“Should Creator partner with other restaurants in the future, these partners would need to hold shared values and principles in ingredient sourcing and producer partnerships,” Vardakostas said. “Sustainability in operations; carbon footprint consciousness; transparency in food preparation; worker growth opportunities and an excellent work environment; and aligned intentions around the guest dining experience.”
Creator is committed to Zero Foodprint, a program founded by restaurateurs Karen Leibowitz and Anthony Myint (Mission Street Chinese Food, The Perennial), in which restaurants conduct an assessment to estimate their carbon footprint, then devote a percentage of sales toward agricultural systems that promote healthy soil development.
The restaurant sources pasture-raised beef from Country Natural Beef, a co-op of family-run cattle ranches. Salad greens are produced by Plenty, an indoor, pesticide-free, vertical farm. Its packaging is compostable.
Staff start at $16 an hour and are entitled to book budgets and paid breaks for reading.
Vardakostas’ position is that Creator’s machines will do the dirty work of repetitive, fast-food cooking so that humans can develop other skills. The belief that automation will free us up to pursue more cerebral activities is a classic argument of technocrats. And in some areas of our lives, like out-of-office email replies, that’s true. But if you believe that manual labor robots are just here to help, without disrupting our labor market, ask an auto worker.
As far as the technological capability for machines like the ones at Creator, we’re about five years away.
“I think for sure at Starbucks, within five years, you could have a robotics system that would be able to replace at least 80 percent of what a barista does,” said Nima Fazeli, an assistant professor with the Robotics Institute and the Mechanical Engineering Department at the University of Michigan, Ann Arbor. “The hard part is where it has to interact with people. The random human element is really hard to predict as a roboticist.”
Fazeli works on inference, modeling, control, and learning algorithms applied to robotic manipulation; in other words, the science and art of robots interacting with the physical world. Fazeli helped build a robot that plays Jenga, not just using visual assessments of tasks to calculate angles, force, and pressure, but with a sense of touch. He said that taking a customer’s coffee order, grinding beans, steaming milk, even spelling your name wrong; these are tasks a machine will do just fine. But what happens when a customer jumps over the counter, or tries to pay with foreign currency, or any number of actions for which engineers have not provided contingencies?
I’m trying to imagine how a robot would react to my wife’s request for a McFlurry with double Oreo, only a little bit of chocolate sauce, and not blended too much. Or movie theater popcorn with “staggered butter,” administered when the bag is half-filled, then again on top.
“The traditional, old-school way of robotics is rule-based systems, in which you have this deterministic set of rules and the robot looks up which rule applies where and executes something,” Fazeli said. “And that works great as long as the engineer who designed those rules is able to anticipate all the things that’s going to happen. The issue with real-world robotic deployment, especially in unstructured environments, where it’s very difficult to anticipate all this variation, is that you just cannot come up with enough rules. A robot eventually gets confused.”
As the technological ability rushes to meet our needs, the biggest challenge is that robots are a very expensive investment for thin-margin businesses like restaurants.
“I don’t think people realize that there’s this huge back-end cost, not only of engineers’ time, but also of R&D that goes into this,” Fazeli said. “Many of these things are possible. It’s just that a lot of them are not economical.”
Like building a wood-burning pizza oven or a marble bar counter, a cooking robot is an upfront cost that has to be amortized over a long period. Creator declined to comment on the total amount invested in R&D, or the timeline or level of growth needed to balance its sunk costs. But it is eager to list the pedigree of its consultants — engineers, roboticists, designers, chefs, and restaurateurs formerly of Tesla, NASA, Apple, Walt Disney Imagineering, Chez Panisse, Momofuku, and Fat Duck. At a minimum, $18 million has been invested in development. That means that expansion is necessary before the $6.07 burger can be validated.
Because of its scale, a company like Amazon can afford to develop robots for its fulfillment centers (what we used to call a warehouse). For independent restaurants, which last 10 years if they’re lucky, it’s unfeasible. For chains, it may be a good long-term investment.
“If you have gifted engineers that are able to do these things, fine,” Fazeli said. “The question is whether that’s economically viable. It depends on the company. For Amazon, it’s worth it because they’re so big that these systems are going to save money for them.”
Rather than buy his technology, the burger industrial complex is more likely to replicate it. Either way, automation in fast food cooking is inevitable. If you can create a machine to cook burgers, you can do it with burritos, ramen, and soy lattes.
No, fast food cooking jobs are not the best jobs. Certainly not the middle-class, secure union jobs of car manufacturing that were displaced by automation and free trade. But they are 3.8 American million jobs. These aren’t the temporary summer jobs once done by teenagers. These positions are more likely filled by recent immigrants and adults who need to support their families. For many of us, these jobs were our first employment experience. Or the first upon arrival in this country. They are some of the most accessible jobs in America. But they pay horribly due to a federal minimum wage that has stagnated at $7.25 since 2009.
And as organizations like Fight For $15 gain momentum in their campaign for livable wages, automation will become an even more appealing alternative to corporate shareholders, compared to their nightmare scenario of paying employees fairly for their work.
However, we don’t have to just wait around for 3.8 million jobs to evaporate.
In a recent paper, Automation and a Changing Economy, Alastair Fitzpayne and Ethan Pollack of The Aspen Institute identified policy initiatives to manage the transition of increased automation.
While the paper doesn’t focus on the fast-food sector, their argument is for the broad application of retraining affected employees through government investment in wage and education subsidies (because it costs less to help people get jobs than to let them slide into poverty); better unemployment payments and services; targeted investment in business development for areas hit hardest by automation; and worker-training tax credit for employers.
The researchers found that between 1996 to 2008, the percentage of U.S. workers receiving employer-sponsored or on-the-job training fell 42 percent and 36 percent, respectively. Suggesting that this will continue to decline, they cite a recent Accenture survey of 1,200 business leaders in which 74 percent said that they intend to use artificial intelligence to automate tasks in their workplace over the next three years, but only 3 percent plan to increase investments in training over the same period.
There is currently legislation, similar to Fitzpayne and Pollack’s recommendations, being proposed to counter the disruption of automation in the U.S. labor market. The Investing in American Workers Act was introduced in the Senate in 2017, followed by a companion bill in the House in 2018.
We won’t lose all 3.8 million jobs in fast-food, and they certainly won’t disappear overnight. But the fast-food sector’s convergence of technological advances and rising wages make an automation disruption inevitable.
None of this detracts from the burgerbot looking cool. And doubtlessly making precise, quality burgers. It just means we need to look past the novelty of a robot cooking a burger so we can have a more robust and realistic conversation about our future.