What's left to teach . . .

. . . as AI comes for us?

When an 18th century Virginia lawyer like George Wythe was looking for a young man to hire as an assistant, a principal requirement was that the young man write legibly and swiftly. Letters and legal documents had to be copied efficiently and without mistakes.

When a 19th century Philadelphia merchant like John Wanamaker was hiring a clerk in his department store, the young man or young woman needed to have a good head for figures. Accounts needed to be kept and checked. Numbers didn't add and subtract themselves.

When a 20th century industrialist like Henry Ford II was hiring engineers, it went without saying that they were quick with a slide rule. No car design was better than the calculations that supported it, and no calculating device was more versatile than the slide rule.

What will employers in the 21st century require of the young people they hire? I think about this a lot, being a teacher of young people who will soon be trying to get hired.

I can tell my students what will not be expected of them. They will not be given a handwriting test. Typewriters and then computers made handwriting obsolete in the workplace.

Their ability to do sums and products will not be tested. Mechanical adding machines rendered that skill unnecessary.

They won't have to show that they can calculate square roots and trigonometric functions on a slide rule. Electronic calculators drove slide rules into museums.

The point is that marketable skills change over time. They are still changing, perhaps faster than ever. The broad embrace of artificial intelligence will render many currently essential skills obsolete.

Computer coding? Even five years ago this seemed like a sure ticket to a promising career. Now up to a third of new code is written by AI. That proportion will grow, eliminating millions of jobs for human coders.

Financial analysis? Almost anything a human analyst can do, AI can do better already or will do better soon.

Medical diagnosis? This is an exercise in pattern recognition, at which AI excels.

Writing? In the professional workplace, where the emphasis is on clarity and timeliness, current versions of AI are better than 90 percent of humans. AI is getting better while humans are not.

The world of literature is different. Many readers value the quirkiness of human writers. Quite possibly many always will.

Yet some of the most successful authors in genres like romance and suspense already subcontract some of their writing. Ghost writers for celebrity memoirs and the like deliberately suppress their own voices. AI can flesh out a plot outline or mimic the style of your favorite celeb as well as a human, and do it faster and cheaper.

Will there be anything left for humans to do?

Definitely, although much of it will be work not usually associated with a college degree. It will be a while before an AI assisted robot can install new brakes on your car. Plumbers’ jobs are safe. Home construction has stubbornly resisted automation, and its stubbornness will continue. Nursing care for the elderly will continue to employ human hands.

The professional and white collar jobs to which college graduates have long aspired are more susceptible to displacement by AI than blue collar and manual jobs. It's called artificial intelligence — not artificial dexterity — for a reason. AI was created to deal with words and images. It’s very good at this. It's not so good at dealing with the physical world. It will get better, but the very physicality involved means that improvements will be incremental rather than exponential.

So what will make my students employable?

First, the ability to communicate orally. This gets short shrift in school but is the principal way people connect in the broader world. It's also something AI cannot do. A talking avatar isn't the same thing as a talking human.

Second, the ability to evaluate sources of information to form the basis for reliable judgments. Theoretically this is not beyond the ability of AI, which could sample various sources and attach weights to them. But practically there will long be questions about whether the AI itself is biased.

Third, the ability to inspire and motivate. Leaders do this by their actions as well as their words. They model commitment and enthusiasm in a way AI can't.

Fourth, the ability to empathize. This is related to the ability to inspire but is broader. A successful leader — a successful person — requires the capacity and willingness to see the world through other people's eyes. This depends on both lived experience and the ability to extrapolate emotionally from that experience. AI has no lived experience and will have a hard time faking it.

What kind of students, then, will be best placed to find jobs?

Not the ones trained in specific skills, for it is in these areas that AI is making the most rapid strides.

Rather, students trained broadly, in what it means to be human. This is what the humanities do.

History students will be particularly well placed. Historians have always had to evaluate sources. Historians have to try to see the world through others’ eyes. Empathy comes from this effort.

We history teachers have to do better helping students learn to communicate orally. But this is simply a matter of will and logistics.

There has never been a bad time to study history. Now is a very good time.