91ÖÆƬ³§ÊÓƵ

In thinking about the jobs of the future and what that means for educators today, there are four general points to keep in mind:

First, there will be jobs in the future. Technology has changed dramatically over thousands of years. And after thousands of years of changes, there are still jobs. It used to take more than 70 percent of Americans to feed us, now it can be done by 2 percent. And yet the others still found employment in new fields that could not have been imagined, such as software engineering and restaurant- and hotel-service-industry jobs. New technologies may make some occupations obsolete, but at the same time, technology raises incomes and creates even more jobs. It is conceivable that our great-grandchildren will live in a world without work but, as the noted artificial intelligence scientist Andrew Ng quipped, worrying about that now is like worrying about overpopulation on Mars.

Second, no one knows what those jobs will be. In the mid-1990s, analysts predicted that a richer population would travel more, generating a need for travel agencies. The travel prediction was right, but the travel-agent prediction was badly wrong: It failed to forecast the explosion of internet-based travel booking that dominates the market today. Today, it is reasonable to expect that an aging population will generate the need for nursing and home-health-care services. But I cannot be sure that advances in AI will not lead to robots with the social abilities and dexterity capable of replacing some humans.

Third, without knowing the precise jobs of the future, we still know the best way to get them: education and more of it. In 2016, 85 percent of people between the ages of 25 and 54, with a college degree, were working, compared with 70 percent with a high school diploma or less. That difference reflects the fact that with more education—such as the attainment of a higher education degree—there is a greater likelihood of entering the labor force and a smaller chance of being unemployed. Current predictions on automation’s impact on the job market conclude that the occupations and skills that are least susceptible to automation tend to be those that require more education. While many aspects of these workforce forecasts will be wrong (see Point No. 2), the broader education trends are likely to hold true.

Finally, schooling should focus more on the skills that complement artificial intelligence rather than those that are substitutes for it. This general rule is easy to enunciate but not well-defined in practice, so its implementation requires continued debate and monitoring. Machines today have difficulty with creativity, expressing empathy, and other soft skills.

It is important that education focus on students’ development of these skills along with the ability to acquire new skills in the future. In a world where machines will become even more important, I share the general enthusiasm for STEM—for science, technology, engineering, and math. But it is worth seriously considering the possibility that some of those tasks, like routine computations, may be better left to machines. The type of STEM matters—teaching the details of Euclidian geometry may have been important centuries ago—but given limited instructional time, understanding how to interpret data through statistics should be a much higher priority.

Some of these are generalities. The future has always been surprising. 91ÖÆƬ³§ÊÓƵ—more of it and better—is the most important step in preparing ourselves to thrive in that uncertainty. Making sure that education will help us build better machines and that we better complement those machines is the key to a successful economic future for our children and society.