Why We Shouldn’t Panic About AI

It sometimes seems that today’s commentators cannot view developments through anything other than a binary lens—each new event or technology will either bring peace and bounty or catastrophe. The debate over artificial intelligence often exemplifies this tendency. On one side, AI seems poised to bring an explosion in productivity, increased leisure, and more freedom from life’s less pleasant tasks. On the other, AI heralds a bleak future of mass unemployment and a dispirited and restive population, in which a few benefit greatly, while others live purposeless lives on a guaranteed government income.

The skeptics paint a dark picture. In a recent article entitled “Something Big is Happening,” AI startup founder Matt Shumer declared that the technology is “coming for” anyone whose job “happens on a screen,” including coders, accountants, writers, journalists, and even medical doctors. Even some AI developers—among them Dario Amodei, Sam Altman, and Elon Musk—see the inevitability of what might be called a jobs apocalypse and accordingly advocate for a universal basic income to care for the soon-to-be unemployable millions.

Of course, AI and robotics will displace many workers, as have all technological revolutions in the past. Maybe the coders of whom Shumer writes will need to apply their keen intelligence and fine talents to other fields. Perhaps superior AI-powered robots will supersede miners, oil field workers, and others in dangerous and dirty jobs as well. The list of potentially obsolete occupations will certainly grow longer as AI develops and will likely include administrative assistants, receptionists, financial analysts, travel agents, junior-level graphic designers, and others.

But in this case, as in all past innovation waves, the new technologies should also create new business and employment opportunities. The efficiencies and cost reductions brought by AI and robotics, even as they supplant workers, will raise real incomes elsewhere in the economy. Accordingly, they will boost demand for more goods and services in ways facilitated by the new technology. Those demands will create more jobs for ousted workers and others. If history is any guide, that job creation will far exceed what prevailed before the technological wave began.

This pattern is well established—from the steam-powered machine revolution that enabled mass production to the electronics revolutions of radio and television to computers and the internet. With each wave, it was easy to worry about the jobs threatened by the new technologies. What has been harder to see early on in each revolution were the business and employment opportunities yet to emerge. A few examples illustrate the case.

During the machine revolution in early nineteenth-century England, the spinning jenny and the mechanical loom could produce more cotton and wool cloth in less time and at less expense than the hand techniques previously used. Those who were displaced famously called themselves “Luddites” and tried to prevent the technological changes by sabotaging the equipment. The use of the machinery nonetheless expanded, efficiencies ensued, and the price of cloth fell so that more people—many of middling incomes—could afford much more clothing than they had before. The textile industry grew to meet the increased demand, ultimately employing far more people than previously, while the expanded wealth from more jobs and greater cost savings enlarged employment prospects and incomes across the whole economy.

Several technological waves later, the introduction of robotics into factories during the 1960s brought a similar panic. The robots replaced assembly line workers and removed the need for highly paid efficiency experts, who ensured that workers made no unnecessary motions in executing their tasks. The fears of the sixties sound familiar to modern ears. President John F. Kennedy warned that automation threatened the “dark menace of industrial dislocation, increasing unemployment, and deepening poverty.” At one point, a group of academics, including several economics and scientific Nobel laureates, issued a report warning of mass joblessness and claiming that automation had severed the once-secure link between jobs and incomes.

In another echo of today’s rhetoric, President Richard Nixon floated the idea of a guaranteed Basic Income to support the anticipated mass of unemployable people. But as happened with textiles over a century earlier, efficiencies and reduced costs created wealth, expanded demand, and, even with the robots, actually raised manufacturing employment almost 25 percent over the course of the decade. It is ironic indeed that those most romantic about factory jobs—typically people who have never worked in a factory, nor desire to do so—now view the 1960s as a golden age of good-paying manufacturing jobs.

With each technological wave, initial worries gave way to efficiencies, cost reductions, new opportunities, and an expanded jobs market, albeit in new and modified occupations. The computerization of business eliminated thousands of clerical positions, but it also enabled employment and profit opportunities in unexpected areas, such as next-day or same-day delivery firms, not feasible before computerization. Today, these firms employ millions, many at higher pay than the displaced clerks once commanded.

Though it’s futile at this early stage of the AI revolution to predict the new opportunities that might arise, some early indications are apparent. Take, for instance, Shumer’s concern for human coders. AI can only recount already existing knowledge; though it can code effectively and efficiently, it offers little help in determining what apps should be coded. That requires judgment drawn from the background knowledge possessed only by seasoned executives and marketing experts. That same need for discernment is even more prominent in other occupations that Shumer wrongly lumps in with coding—law, medicine, writing, and areas that deal with people. Similarly, AI may enable self-driving school buses, but these vehicles will still need an adult to watch the kids. Robots cannot replace police or game wardens, therapists, or personal life coaches. They can replace people who make plumbing fixtures and electrical equipment but not the plumbers or electricians whose judgment will still be needed.

The mass unemployment often anticipated is likely not to materialize to such a dramatic degree because AI-driven cost reductions will encourage risk-taking, along with new opportunities and prospective employers that might otherwise never have existed.

Consider how AI technologies have already made media much less capital intensive. What required a studio full of specialists and expensive equipment only a few years ago can now be done with a laptop in an apartment. While this innovation may have terminated many legacy media jobs, it also opened opportunities and expanded employment for numerous independent operators in broadcast, writing, and journalism.

This is far from a complete list of AI-related opportunities and sources of employment just beginning to emerge. No doubt many surprises are in store. After all, when James Watt made the steam engine practical, he saw it as a way to pump water out of coal mines—he had no notion of locomotives or steam ships. Two and half centuries later, the inventors of smart phones and GPS never anticipated Uber.

It is, of course, possible that things will be different this time, that the patterns of the past will no longer apply. But that seems unlikely. In any case, it will take a long time before the effects of AI are fully realized. The technology may move at lightning speed, but complete adaptation will require changes in business and government organization as well as everyday practice. Adaptation will also demand adjustments in regulations and legal structures. These matters move much slower than lightning. It took 20 years after the invention of the personal computer to find these devices on every office desk and fully integrated into the operations of business and government. AI adoption may move faster than that, but its effects—good, bad, and indifferent—should unfold at a pace that will allow people and practices to adjust.