AI Policy Congress – Part 6 Manufacturing & Labor

2019-02-25 - 6 minutes read

Written by Willie Boag

Technology’s Past, Present, and Future Impact on Work

Is the diffusion of AI going to harm middle-class jobs? Broadly speaking, we can expect to see labor displacement (shorter-term) and labor adaptation (longer-term). Some experts say that in the coming decades, technology will likely change the type of work that is done, but AI won’t be powerful enough to truly replace humans for most jobs. Although a lot of the tech is new, the policy questions are ones we have been wrestling with for decades. During a session moderated by former White House Chief of Staff Denis McDonough, the panel discussed the past, present, and future of technology’s impact on work.

This is not the first time in American history that technological innovations unsettled the dynamics of the labor market. As panelist Philip A. Miscimarra, Partner at Morgan Lewis and former Chairman of the National Labor Relations Board (NLRB), explained, one of the leading historical NLRB cases is one in which the invention and production of prefabricated doors disrupted the 1960s-era expert labor of door craftsmen. The market adapted, and those who were displaced found new jobs in which they could apply their skills.

The same lesson, Miscimarra suggested, will likely hold for jobs that AI displaces now. Although some constituencies might resist tech disruption of their jobs, Miscimarra reminded the audience that throughout history, opponents have tried to delay or slow down technical adoption, but they never stopped it completely. He argued that policy barriers to prop up some jobs may not work because “in a global economy, to the extent that particular countries erect barriers to technological change – artificial intelligence is one example – other countries will not erect the same barriers.”

Denis McDonough, Philip A. Miscimarra, Erik Brynjolfsson, and Jacqui Canney during the Manufacturing & Labor panel. Photo by Leilani Gilpin.

As we look forward, we might wonder what kind of displacement AI will bring about. MIT Professor Erik Brynjolfsson pointed to his research with Tom Mitchell, which used a 23-question rubric to determine whether a machine can replace a job. The takeaway: AI will replace tasks, not jobs.

There has been much talk recently about radiologists’ days being numbered because of breakthroughs in computer vision, but according to Brynjolfsson, a radiologist’s job isn’t just interpreting images: the job can be broken down into 26 tasks, including communicating with patients and other medical professionals. Generally speaking, of the 900+ jobs that Brynjolfsson investigated, he didn’t find any that lent themselves completely to automation. Instead, he argues, we can expect a future of machines helping humans better perform their jobs.

Jacqui Canney, Executive VP of People Operations at Walmart (the largest private employer in the U.S.), agreed. Currently, Walmart is using AI to automate simple tasks, such as unloading shipping trucks and cleaning floors, in order to create more time for their associates to interact with and assist customers. This shift to more human-centered service allows Walmart to look for associates with a different set of skills, now including more emphasis on emotional intelligence, communication, and digital literacy. Walmart has also invested heavily in skills training for its employees, with help from companies innovating in education sector.

Panelists at the MIT AI Policy Congress discuss manufacturing and labor. Photo by Caty Fairclough.

Armed with this perspective and context, what are we doing to prepare for the future of work? AI will be able to automate many mechanical tasks, leaving interactions and communication to humans. This will create more demand for workers with skills like empathy, which our markets have failed to value throughout history. At least one audience member was skeptical, noting that “people have been saying teachers are underpaid for the entire history of the United States” with little being done to actually retain that human labor. All predicted that these issues will continue to engage academia, industry, and policymakers, and they were right: just weeks after experts at the AI Policy Congress called for more Federal government attention to the future of manufacturing and labor, President Trump signed an Executive Order on Maintaining American Leadership in Artificial Intelligence — with a particular focus on workforce and educational development.

Learn More About the MIT AI Policy Congress

This post is part of a series on the first MIT AI Policy Congress, edited by Grace Abuhamad. Read the rest of the series on the IPRI Blog.