When automated self-checkout lanes started becoming more popular, I remember wondering whether cashiers would start losing their jobs to the new technology. Now, with Amazon Go showing the world that you can walk in and out of a store, products in hand, no line necessary, we see that reality hurtling towards us at a higher-than-ever speed. The White House can keep convening task forces claiming to “put workers first,” but the reality is that 3.5 million cashier jobs alone are at stake.
At a conservative estimate, an increase in the robotics workforce—one way to measure job automation—means that an average of a quarter of a million people will lose their jobs by 2020. Here’s another way to think about it: For every robot per 1000 workers, the unemployment rate increases by 0.18–0.34 percent, and wages for human workers drop by 0.25–0.5 percent. In Trump’s much-applauded campaign promise to bring back American jobs to industries like coal mining, he neglected to mention that these jobs no longer exist—they’ve become mechanized, from self-driving trucks to automated drills. It’s jobs like retail, ones that still employ millions of people, that are really at risk.
Still, these numbers might not seem alarming. After all, the U.S. unemployment rate in April 2018 was 4.1 percent, or 6.59 million people—which sounds charmingly low until you look more closely. For Black people, the unemployment rate is 6.9 percent; for Latinx, 5.1. For people without a high-school diploma, it’s 5.5 percent. And unsurprisingly but still devastatingly, for people with disabilities, that number is 8.2 percent. What this means is that the people who stand to be most affected by robot replacement will be those already marginalized in some way.
As it is, the tech industry’s innovations have led to job decline, even if it’s not fully mechanized labor. One example is how smartphone technology has enabled rideshare programs like Uber and Lyft to exist. Such programs have famously bankrupted the taxi industry, which means immigrant laborers are at risk for further displacement. My grandfather, once the head of the Abadan police force in Iran, has been a taxi driver in the Washington, D.C. area for the last 30 years, and rideshare apps curtailed his driving even before age could.
One place we might see mass human replacement is in the food industry, where Latinx and Black workers are almost always the core of so-called back-of-the-house jobs like dishwashing. Mechanizing these positions would therefore push low-wage earners to look for similar jobs that may no longer exist. Furthermore, some of these positions, such as in the fast food industry, may be the only opportunity for these workers to have access to health insurance, time off, sick leave, or other benefits, pushing them further to the margins.
As a former scientist, I’m not worried about robots taking over the menial lab tasks often relegated to techs or undergraduates. I have an advanced degree; perhaps one day our AI will be advanced enough to think through the necessary steps for a sound experimental design, but that day probably won’t be here anytime soon. I am worried, however, that workers who are already at a disadvantage will lose more jobs to mechanized labor. Our society already makes it difficult for people of color and people with disabilities to get higher education or climb the corporate ladder, leaving them jobs that will be the first to go with increased automation.
That said, automation isn’t feasible for every company. My friend Terry Benton is an industrial engineer who has spent the last 10 years in a variety of industries, often working with companies to automate processes. He’s currently the Director of Manufacturing for Carbice Corporation, a tech startup in Atlanta, GA. He told me that, “[while] automation is a ‘new and shiny’ concept, and impressive, it doesn’t always make sense for businesses to fully automate. … [The] harsh reality is that automating existing processes is very expensive and most manufacturing companies who could benefit from it don’t have the capital available to invest in it. I’ve seen tech companies utilize automation to great effect, such as automating simulation, testing, and prototype development, which greatly improves speed-to-market and accuracy and accuracy of design.” One example of a company with that starting capital is Uber, which has been working on self-driving vehicles with decidedly mixed results.
I thought that Terry, as a Black man, might have thoughts on whether marginalized people are in as precarious a position as the numbers seem to suggest. Short answer: Yes. “[It’s] been my experience that the industries that pay the lowest production wages typically employ more people of color,” he said. “I did several projects for a food and beverage manufacturer where the majority of their production workforce were people of color (primarily Black and Hispanic). On one such occasion, we couldn’t justify reducing enough headcount to support automating several processes because their wages were too low (although still legal).
The people who stand to be most affected by robot replacement will be those already marginalized in some way.
On the other hand, I did a project for an aerospace and defense contractor, which employed skilled aircraft maintenance technicians and the majority of the workforce was white and I was absolutely blown away once I found out what their fully-loaded costs were, which was the highest I’d seen of any operation or business I’d been in to-date. … I believe this points to a much larger issue related to centuries-old systems of oppression that have severely limited opportunities for people of color to receive the same education and/or training as white people, thereby often forcing them onto the lowest rungs of the corporate ladder.”
And yet Terry still believes that automation isn’t necessarily a bad thing. “As a Black man whose parents both worked in manufacturing, I support automation… [Rather than] stifling our technological advancement as a people, we should be putting additional effort into improving access to education and training that will provide alternative jobs and careers for people of color when displaced by automation.” On the other hand, even as reports vary on the exact number, there’s no doubt that millions of jobs are at risk across a variety of sectors.
Obviously, dismantling an entire system of oppression would be the best solution for this and countless other inequalities. But large-scale changes of that kind have consistently eluded labor policies, as evidenced by not just the higher unemployment numbers many people of color face but also the systemic violence inflicted upon us. Among these are not just barriers to workplace entry like unpaid internship experience but higher and more severe incarceration rates, police brutality, city segregation, and the closure and decline of local public schools.
I wonder, then, if the solution has to begin at the company level, with upper management taking more responsibility for the livelihoods of its workers—especially the marginalized ones. Automation may not be new, but as technology prices drop and advancement continues, the face of it changes. And as we grapple with how increased automation will affect human labor, it’s vital that we go beyond the overall impact and focus on the workers who stand to lose the most.