This summer, Dishcraft released a new dishwashing robot. It’s only the latest invention that we’ll use in a way that threatens what I’m calling “on-ramp” jobs.
I loved being a dishwasher. Every summer for a few years in high school, I’d get up at 7 and walk half a mile from my house to a small breakfast joint. Don ran the place, Al was the cook, and Rick was backup. Compared to theirs, my job was dead simple: Take the dishes out of the bus trays, scrub them down, put them in a heavy-duty plastic rack, slide that into the dishwasher, wait until the cycle finishes, take out the cleaned tray, put the dishes away on a rack. And repeat—maybe 50 or 60 times a shift. Meanwhile I learned how to care for the machine, I enjoyed figuring out ways to do my job with fewer steps, I got to know Al and Rick, I heard their war stories, and I got a bit of an intuitive understanding of how our kitchen ran. That’s when Al asked me to step behind the oven.
Matt Beane (@mattbeane) is an assistant professor of technology management at UC Santa Barbara and a research affiliate at MIT’s Institute for the Digital Economy.
On-ramp jobs like these (page, research assistant, receptionist) allow us to do something simple and useful in the middle of the action, get paid, build relationships, get a broader view on organizational operations, and make an educated decision about where to go next. My research shows that intelligent machines are sharpening a scalpel that will allow us to carve up work processes in the name of productivity, and that routine jobs are the first to go. Dozens of studies show automation’s looming effects in contexts as diverse as investment banking, policing, education, and internet startups. What’s going to happen to on-ramp jobs in this new world we’re creating?
Dishcraft’s system is a good example to think with. For large-scale dining operations, the company promises that its robot will allow for a significant reduction in the labor required to clean plates and bowls—which still is most of the work for those who do dishes. In a statement, Dishcraft founder and CEO Linda Pouliot wrote that the company’s “vision is to improve the role of dishwasher rather than replace it … Dishcraft is automating the dirty, dull, dangerous part of dishwashing, so it is a cleaner, safer, rewarding job that more people want to do.” The system reliably takes in a stacked tray of about 50 dishes, handles and cleans them internally, and sends them out to be reused. Like many robotic systems, it works because engineers made the job a lot more predictable. The ceramic plates and bowls are custom-built with stainless steel installed in them that allow a robotic scrubber to use magnets to quickly remove all debris. Then the dish is rotated to an “inspection station.” If Dischcraft’s AI vision algorithms determines it to be clean, it will move on to the next stage; if the AI deems it dirty, it will be cleaned again.
That “making the job more predictable” bit extends well beyond the focal washing task. To take advantage of the system, spaces and other equipment need to be reconfigured to make room for the robot and for cart transit routes, workers have to learn to perform more specialized, repetitive tasks like loading and unloading the carts and the robot, assignments need to shift to balance and connect these more granular jobs throughout the facility, and new policies and safety procedures have to be instituted. Decades of research decisively show the same pattern: Organizations have to expend massive effort overhauling preexisting facilities, workflows, and norms just to get new forms of automation to function, let alone deliver a satisfactory return.
Whether it involves thousand-pound, million-dollar robots or intangible, freemium software, new automation is going to mean fewer on-ramp jobs in cities. Business reconfiguration expenditures will be worth it only for large businesses that have many staff and highly standardized processes, which are overwhelmingly located in urban areas. Small businesses don’t have the volume to justify this kind of investment, and there’s too much variation in the work (e.g., handling different types of plates, bowls, and cutlery differently) for this generation of technologies to handle.
social experiment by Livio Acerbo #greengroundit #wired https://www.wired.com/story/killing-dead-end-jobs-only-hurts-us