If you work at FedEx–or really anywhere in the logistics industry–you might want to scan a recent story in the MIT Technology Review. Headlined “FedEx Bets on Automation as It Prepares to Fend Off Uber and Amazon,” it describes how the company is investigating AI-enabled assistants, semi-autonomous trucks, and even “FedEx courier robots” that make their own deliveries (such machines already exist).
The article doesn’t discuss the implications for FedEx’s 300,000-strong workforce. But we can be fairly sure that driverless vehicles and robo-deliveries won’t lead to improved job opportunities for them. Most technology, after all, is designed to be “labor-saving” one way or another, and we can be near-certain the next wave of robots, AI, and machine learning will save us a lot of labor. Logistics in fact could be one of the first industries to be automated, at least partially, research suggests. It involves lots of repetitive processes ripe for technological intervention. Margins are tight and competition is fierce. If FedEx doesn’t do it, then its competitors will.
The question as a FedEx employee, though, is what exactly are you supposed to do if you’re worried? Learn new skills? Okay, but which ones? What skills might complement ones that you already have, and how long might those skills last you before the next automation wave? Futurists generally point to the need for skills that robots don’t have, including empathy, an ability to work in teams, manual dexterity, and critical thinking. But that doesn’t necessarily add up to helpful career advice. It’s too general.
There’s been a flurry of recent reports and research predicting future job losses from automation. But according to Muriel Clauson, an organizational psychologist at the University of Georgia, most of these aren’t particularly accurate and certainly not that useful if you’re trying to get ahead of the automation curve. “So you need to concentrate on manual dexterity? What do I need to do with that? Practice crochet every day?” she says in an interview with Co.Exist.
While studying at Singularity University, in California, last year, Clauson and three colleagues (Jenny Appel, Pablo Orduña, and Laurent Boinot) came up with a possible solution: software that tracks changes in employment and helps workers adapt to what could be coming next. Clauson, Appel, and Orduña then formed a company to develop the idea–what Clauson calls an “AI coach for the unemployed.”
The trouble was, the underlying data to make predictions about unemployment wasn’t “granular” enough, Clauson says, and so the company, named udexter, has now somewhat switched focus. Most technological unemployment research, including a widely cited study from Oxford University, uses O*NET OnLine, “a database of occupational requirements and worker attributes.” She points out that the underlying classification system was developed in 1954 and uses “a taxonomy of skills that’s very antiquated.”
“Job categories are grouped in ways that don’t always reflect the reality of work,” she says. For example, the page covering computer coding jams together system administrators and front-end web developers–very different roles. In considering how technology may affect the workplace, she argues we need a higher level of detail and specificity.
Clauson is therefore working on an alternative to O*NET. It’s based on a system of skill “stacks,” where the skills required for each occupation appear across multiple areas, showing how the same abilities could be useful across industries and activities. That differs from O*NET, which lists single occupations and a checklist of skills required to do that job only.
Once finished, she hopes the new system, which uses crowdsourced data, will serve as a basis for the original udexter software idea, which will allow people to see how their skills (and any future skills they might acquire) could be repurposed. At the same time, udexter wants to be a hub for solutions and policies to technological unemployment and for the people who develop them. That could include “evidence based” education policy and, presumably, the idea of a basic income guarantee.
Predicting the future of work isn’t only useful for individuals caught in the swell. Governments also need a better handle on the shape of the job market, McKinsey and others have said. At the moment, it seems, we’ve barely begun to map the actual state of the modern workplace, let alone what technology might do to it in the future and what jobs might still be available for human beings.