Today we need a lifelong learning movement suitable to an AI-augmented economy.

Originally published on Forbes.com

Millions of students earned their high school diplomas this month. And as the caps and gowns are packed away, another generation will begin to look toward their careers. Most will spend some time in post-secondary education, but the majority will not earn a four-year degree. So even at this time of celebration, many graduates are worried. They hear technology is killing jobs. Even if you can find work today, who knows if it will be around for a full career?

What if these grads found out that a quarter of all of the jobs they know today will disappear in their lifetimes? 

On the one hand, it could be cataclysmic to millions of displaced workers. How would we cope?

On the other hand, it already happened—nearly a century ago. And we did cope. In fact, we thrived.

The biggest historical threat to American jobs isn’t artificial intelligence—it was the tractor. 

From 1910 to 1960, the agriculture industry went from one-third of all U.S. jobs to just 8%, shifting a quarter of the labor force out of farming over the course of one working lifetime. Back then, it meant 9.7 million farm jobs lost; today it would mean some 40 million. That’s the equivalent of all the jobs in the retail, transportation, manufacturing and financial services industries being replaced by new industries. It sounds unbelievable, even insurmountable. But it happened, and we surmounted. 

This transformation could have spelled economic disaster, but instead, the nation’s collective response paved a way to the strongest middle class the world had ever known. Rather than allow a displaced generation to flounder, the American High School Movement was born.

As economists Claudia Goldin and Larry Katz describe in The Race Between Education and Technology, public high schools prepared students for the careers of the day in factories or offices. The pace of agricultural mechanization was so rapid that men and women who came from long-time farming families could see there would not be enough agriculture work for their children. So when companies advertised jobs in new fields which required a high school degree, communities demanded—and funded—the public high schools needed to provide them. 

The American high school movement was a wild success. Few Americans in the early 1900s were educated past 8th grade. In three decades, however, high school enrollment shot up, going from under 10% in 1910 to more than seven of 10 U.S. teenagers attending high school by 1940. For much of the century, mass learning was strongly tied to higher earnings—having a high school diploma put a career-path, middle-class job well within reach for millions. 

But for the students who graduate this month, that’s no longer true. More than 100 million working adults without four-year college degrees are locked out of career-path jobs in part because, whatever skills they possess, they lack the pedigree so many employers now demand.  

Our motivation to reinvent learning-to-earning transitions should be as urgent as it was a century ago. Although we may not see job destruction on as large a scale, expert analyses suggest that 30% of workers’ tasks may be transformed by artificial intelligence (AI) by 2030. Meeting this challenge will demand a massive level of continuous re-skilling, for which our talent management practices and institutions are poorly equipped. In much the same way the high school movement prepared young Americans for good jobs in an emerging industrial economy, today we need a lifelong learning movement suitable to an AI-augmented economy.

Beyond its “jobs of the future” motivation, the high school movement had two more keys to success: a widely-accepted model (“high school”) and a source of money (local property taxes).  

Our economy is more complex than a century ago, so we will likely need more than one solution. At a high level, three competing lifelong-learning models have begun to emerge:

1. College for all. Since A Nation at Risk was published in 1983, U.S. high schools have increasingly been treated as a stepping stone to a 4-year college. A Millennial generation that is America’s most educated ever also has unprecedented student debt and stagnant median earnings. Even worse, barely half of those who start college complete a degree. Competency-based colleges like Western Governors University hold promise for adult working learners, but so far, online colleges have boosted debt more than incomes.

2. Apprenticeships and work-based learning. Increasingly, U.S. businesses, industries and policymakers have looked to adapt apprenticeship practices from northern Europe, in partnership with a variety of labor unions, community colleges and specialized outsourcers. Tennessee’s Centers for Applied Technology are a great example, or IBM’s P-TECH high schools, which are industry-led, publicly-funded programs in Brooklyn and eight states. 

3. Last-mile training/re-skilling on demand. More recently, a plethora of “last mile” job training programs and curricula—typically between 3 and 12 months in length—have emerged. These include “coding bootcamps,” online courses from Udacity or Coursera and non-profit programs such as EdX and Merit America. There are also new training/placement on-ramps for young adults, veterans and other groups, such as Per Scholas, NPower and Last Mile. AI should play a positive role in this model, allowing real-time “fill-ups” of needed skills.

Money was crucial in the rapid adoption of the high school model. It had a clear way to gather and oversee funds: call a town meeting, elect a school board and vote for local property taxes. 

How will we fund lifelong learning? 

Our current approach to postsecondary education puts most of the financial risk on students, most of whom are ill-suited to judge the returns in advance. One solution gaining popularity is to make either two or four years of public college free or debt-free, paid for by state or federal taxes (as high school was locally funded). A second approach is to re-engage companies as the source of training dollars, perhaps by changing our tax code and accounting standards so there are better incentives to invest in human capital, similar to those for R&D or capital equipment. 

Overall, however, we should apply more creativity to this challenge. For instance, what if our public policies supplemented people’s income for, say, six to 18 months while they trained for new jobs? It’s the idea of “basic income to rise,” as Gene Sperling has called it. Meanwhile, why can’t a mix of policy changes and social innovation create new loan standards which treat financing of last-mile vocational training more like a small business loan to be paid out of higher earnings—rather than like a consumer loan subject to FICO scores? It is, after all, an income-enhancing investment that benefits our communities with higher-skilled workers.

While there are good reasons to doubt that the workforce displacements ahead of us will be as radical as those which inspired the American high school movement a century ago, we should pursue a lifelong learning system with a sense of urgency. Still, we should worry less about the pace of technological change, and more about whether our institutions are up to the task of using technology and innovation to augment people’s talent. We will only be able to re-connect learning to earning if our education, politics and labor and capital markets rise to the challenge. 

As a society, we owe our newest crop of graduates—and every hard working American—no less.