AI adoption, labor shortage accelerate need for upskilling

Nuro Inc., a company that makes autonomous delivery vehicles for goods such as pizza, relies on AI and specialized technical skills to build and maintain its fleet. How it fills its labor requirements illustrates one of the primary ways that AI is affecting the workforce and upskilling.

Nuro's self-driving delivery vehicles ostensibly reduce the need for delivery workers. At the same time, the company, headquartered in Mountain View, Calif., with a manufacturing plant in Las Vegas, is creating new jobs that require a higher level of skills.

Even though AI technologies can potentially eliminate jobs, that's not been the case. Unemployment remains at record lows, and job openings are still high. The greater problem facing the workforce is in upskilling employees to tackle new types of jobs that AI has created, according to some of the panelists at a recent Brookings Institution forum on AI and upskilling.

The underpinnings of Nuro's work are software and AI learning models that autonomous vehicle systems use for perception, prediction and planning. About 30% of the engineers who build the technologies are on visas because of a labor shortage in the U.S., said panelist David Estrada, chief legal and policy officer at Nuro.

But Nuro is also training potential job candidates in the U.S. Last year, in cooperation with a community college, it launched a course to train people as vehicle technicians, enabling them to "gain the skills necessary to be hired by us," Estrada said.

Economists measure the amount of automation in the economy through labor productivity as a proxy for automation. The more work is automated, the more an employee can accomplish in an hour on the job, said panelist Gad Levanon, chief economist at the Burning Glass Institute. The institute, a nonprofit, was established by Emsi Burning Glass, a labor analytics firm.

But productivity growth numbers are disappointing, according to Levanon. "In the decade before the pandemic, we had the weakest labor productivity in U.S. history," he said. There were hopes that the adoption of digital technologies driven by the COVID-19 pandemic would provide a bump in productivity growth, "but so far, we haven't seen it," he added.

It seems sort of paradoxical that occupations don't disappear, and yet so much changes.

The main reason for the lack of productivity gains is the increasing difficulty companies are having with fully replacing workers with technology, Levanon said. Jobs that could be completely replaced with technology were automated more than a decade ago, he said.

But an absence of productivity gains doesn't mean that technology isn't bringing swift changes, said panelist Morgan Frank, assistant professor at the University of Pittsburgh School of Computing and Information. "It seems sort of paradoxical that occupations don't disappear, and yet so much changes," he said.

Rapid technological changes "can still create costs even if [they don't] create unemployment," Frank said. An example might be a worker who struggles to keep up with in-demand skills. It's a constant reskilling and upskilling effort, he said.

Levanon said skills are changing rapidly in business-related jobs such as marketing, sales and HR. The underlying systems for these positions are becoming data-centric, "but to make good use of that data, there is a need for people in those departments who know how to use it," he said.

Another trend is that the number of job advertisements requiring a bachelor's degree has been declining, according to Levanon. He said wage growth for jobs that don't require a bachelor's degree has been much stronger than for jobs that require a degree.

Levanon said dropping the bachelor's requirement is, in some cases, a response to the labor shortage.

"When you are desperate to get someone, you are willing to compromise on the requirements," he said.

Patrick Thibodeau covers HCM and ERP technologies for TechTarget Editorial. He's worked for more than two decades as an enterprise IT reporter.