Artificial intelligence offers opportunities for engineers, and according to Kai-Fu Lee, chairman and chief executive of Sinovation Ventures, writing in The New York Times, it does not present a near-term prospect of a sci-fi dystopia. What it will do, he writes, is outperform humans at many tasks—such as an AI-based “spreadsheet on steroids” evaluating loan-repayment histories. As AI spreads across many domains, it will eliminate many jobs. Retraining may help people find new jobs, but Lee expects few laid-off production line workers to become trial lawyers or other highly paid professionals requiring creativity and cross-domain thinking. They may more easily acquire, or already have, the people skills necessary to become social workers, bartenders, or concierges, he writes, but then asks, “How many bartenders does a society really need?”
Lee’s outlook is grim. He says that large chunks of money created by companies that develop and deploy AI will have to be transferred to displaced workers through Keynesian policies of taxation and increased government spending.
But before we reach that stage, it’s worth a look at how else displaced workers might be helped. Researchers Martha Ross and Natalie Holmes at Brookings have studied 78.9 million adults aged 25 to 64 in 130 locations. They identified 11.3 million who were out of work but might be interested in or benefit from workforce development assistance. (They excluded people receiving retirement or disability benefits, most students, and likely stay-at-home parents with employed spouses.)
They identified eight programs that might help, including bridge programs to prepare people for further education and training, transitional jobs programs that provide short-term subsidized employment, social enterprises that combine nonprofit and market-driven approaches, job-search assistance and counseling through job centers and other employment programs, sector initiatives that identify workforce needs in a given industry and location, two-generation programs that link job training for parents with early childhood education for their children, and apprenticeships that combine paid employment with on-the-job training. The eighth is Accelerated Study in Associate Programs (ASAP), designed by the City University of New York to increase the graduation rate of low-income community college students.
Ross and Holmes evaluated the applicability of these programs across out-of-work populations of varying ages, levels of education, and earnings history. Young, less educated, and diverse people could benefit from all eight programs, they write. Less educated prime-age people could benefit from all but the ASAP program. And motivated and moderately educated younger people could benefit from all but transitional jobs and social enterprises.
Many groups, they find, can benefit from job-search assistance and counseling, including less-educated people nearing retirement, moderately educated older people, highly educated and engaged younger people, and highly educated older people.
If Lee at Sinovation Ventures is right, these programs will be of limited value. AI will not convert artisans into assembly-line workers, as in the industrial revolution, or typists into computer operators, as in the computer revolution. Governments will pay displaced workers to perform what he calls “service jobs of love”—such as accompanying an older person to a doctor’s appointment or mentoring children. He explains that the volunteer service jobs of today may turn into the real jobs of the future.
However, the programs identified by the Brookings researchers are worth exploring, especially by local leaders who can, as Ross puts it, “… better understand who in their community wants or needs work and which strategies are best suited for connecting their diverse out-of-work residents to employment.”
Visit my blog for links to Lee’s article and to an interactive version of the Brookings report.