Flippy and friends: should we care if intelligent robots take our McJobs?
Image credit: Miso Robotics
A major US fast-food chain has announced that artificially intelligent burger-flipping robots will be rolled out in 50 of their restaurants over the next two years. What are the limits of what they can do, and could they improve working conditions in kitchens?
Kathryn’s colleague has been in a sour mood for weeks: he recently cut off his finger while hurrying to prepare food for customers at the London restaurant where they both work. It is under these conditions that the chefs begin to ponder the benefits of sharing their work with robotic colleagues.
“I would have no problem working with a robot as long as it was friendly,” Kathryn said, thoughtfully. “Language abilities would be a definite bonus.”
More than ever before, we are debating whether robots will take our jobs. The consensus is: yes, they will, but it may take a while and some jobs will remain which are inherently human. This centuries-old fear of technological unemployment has been rekindled by the rapid expansion of artificial intelligence (AI) and robotics in the workplace.
Robots filling in as receptionists, monks and police officers may be the most headline-grabbing examples of robots at work, but roboticists are primarily focused on the development of machines to perform the most “dull, dirty or dangerous” jobs.
These machines vary from weaving machines to cutting-edge search-and-rescue robots, which crawl through the ruins of disasters searching for survivors.
Of all the unstimulating jobs, none is as disparaged as burger flipping. Dull, dirty, sometimes dangerous, jobs in fast-food restaurants were nicknamed “McJobs” by a much-quoted Washington Post article, “McJobs are bad for kids”, which emphasised the poor wages and career prospects associated with this work.
“[Our] employees don’t enjoy working on the grill. It’s hot, greasy and hard on their wrists,” said John Miller, chairman of Cali Group, the parent company of US fast-food chain CaliBurger. “As a result, we have a very high turnover on the grill position.”
It is good news, then, that from early 2018, dozens of CaliBurger restaurants in California will begin employing someone who “loves” their job flipping burgers.
These restaurants – described by Miller as “proving ground” for new technologies – will be the first in the world to introduce an AI-enabled food-preparation robot in their kitchens.
The robot, named “Flippy”, was created by a California-based tech company, Miso Robotics, with funding from Cali Group. Flippy takes the form of a robotic arm and sensor bar on a small wheeled cart. It can be installed by a standard grill or fryer in minutes.
Flippy is equipped with thermal, 3D and vision sensors, allowing it to monitor temperatures across the grill and detect when a patty is ready to be flipped or plopped onto a bun. It sends an alert for human assistance to add cheese, condiments and other toppings to a cooked burger.
The automation of kitchen chores is nothing new – we think nothing of using dishwashers or blenders – and as early as 2012 Momentum Machines unveiled a robot capable of cooking 400 burgers an hour.
Flippy is noteworthy as a kitchen robot because it embodies AI. So far, no other company has created a kitchen robot capable of learning new skills.
Machine learning allows a computer to identify patterns based on masses of previous data, rather than being pre-programmed to perform tasks. With each flip of a patty on the grill, correct technique is reinforced and poor technique is discouraged. This allows the robot to refine its skills and even learn to complete new tasks.
“We’ve got an adaptable piece of hardware that operates with our AI software, Miso AI, which gives Flippy eyes and a brain,” David Zito, CEO of Miso Robotics, told E&T. “Much like children learn new skills, Flippy is equipped with the ability to learn from its surroundings and improve, as well as expand its skillset.”
While Flippy will initially work just at flipping burgers, its creators and employers hope that before long it will learn to grill chicken, chop vegetables and detect pathogens in food before it is served to the public.
In theory, Flippy could become a diligent, tireless commis chef. With sufficient training, it would be capable of learning countless new tasks, adapting to seasonal menu changes through the year. Zito hopes that it will eventually be able to accommodate a restaurant’s entire menu.
It is this embodiment of AI within robots that induces 21st-century panic about technological unemployment. A robot equipped with eyes and a brain for learning new skills poses far more of a threat in the workplace than a machine only capable of performing limited, repetitive tasks.
These “double threats” are most concerning for low-skilled workers such as fast-food kitchen staff, who are already engaged in struggles to improve their generally poor wages and working conditions. In California, where Flippy is due to begin working in 50 CaliBurger restaurants over the next two years, the state government is planning to raise the minimum wage to $15/hour (£11.20/hour). This is significantly higher than the $11/hour (£8.20/hour) the typical US chef earns, according to the Bureau of Labour Statistics.
With the cost of kitchen labour due to rise and consumers demanding faster service, it is possible that soon other fast-food chains will begin to look at the possible robotic additions to – or alternatives to – human labour in their kitchen.
“The industry has remained largely unchanged over the last thirty years and restaurant operations continue to struggle to keep up with consumer demands and expectations,” said Zito. Zito and Miller hope for a “co-botic” work environment, in which robotic assistants take over dirty, boring, time-consuming tasks in the kitchen, reducing the burden on human chefs.
This would not be such a dramatic change from the way that a lot of tasks around the kitchen are already partially automated, Kathryn said.
“I wouldn’t mind some of the tasks that are already pretty automated being further automated,” she said. “Quite a few of the tasks that we do just involve plopping something into an automated deep fryer or microwave and waiting for it to drop back out again, which obviously involves minimal skill.”
While the artificially intelligent Flippy poses a greater threat to chefs like Kathryn than anything that has come before, the possible rise of “co-botic” kitchens could preserve jobs while reducing workload.
A less optimistic – but arguably more probable – view is that the introduction of Flippy and other robots will not change chefs’ workloads: some McJobs will be taken by robots, while decimated teams of chefs could be forced to work harder at the tasks which cannot be automated.
Some senior roles in the kitchen remain far beyond the abilities of Flippy, however - most obviously creating recipes to be followed by human chefs and robotic assistants.
Previous attempts to use AIs to create recipes have ended disastrously. A neural network trained on 30,000 recipes from books creatively - but tastelessly - suggested dishes such as “Banana Washed” and “Beef Soup With Swamp Peef and Cheese”. IBM’s “Chef Watson” has fared only slightly better.
“It will never be able to add that creative spark necessary for planning a delicious, mouth-watering menu”, said Zito. “Taste is an important part of the dining experience, so a chef’s expertise will always be invaluable.”
Intelligent robots such as Flippy are already coming for the McJobs and that is unavoidable: both Zito and Miller consider this technology to be important in the future of the industry.
“We believe AI-driven robots will become a standard for the quick-service restaurant industry,” said Miller. “We think that this transition will be analogous to how the computer replaced the typewriter in offices.”
Whether this will worsen or improve working conditions for kitchen staff is a matter for restaurant managers, who must cynically weigh the human touch of restaurant staff against their rising wages.