Artificial intelligence is not going to take all our jobs

Why isn’t everyone already unemployed? After all, experts have been predicting for decades that technological automation might soon take most of our jobs. With the rise of even more powerful artificial intelligence (AI) and robotic technologies, these same fears are leading to a fresh round of worse-case forecasts about the displacement of skills and sectors. 

A new study forecasts that “up to 49 percent of workers could have half or more of their tasks exposed” to AI-powered large language models like OpenAI’s GPT-4. A decade ago, researchers at the University of Oxford published a widely cited study that similarly predicted 47 percent of U.S. jobs were at high risk of automation. 

The good news is that the sky is not falling due to AI. Largely the opposite of what the Oxford scholars predicted about AI job losses came true. Since 2013, the U.S. economy has added 16 million jobs and the unemployment rate fell steadily despite continued automation and increasing robotization of many workplaces. The profession that the Oxford report said faced the highest risk of technological disruption—insurance underwriters—saw employment grow 16.4 percent. Meanwhile, the biggest employment problem the economy faces today is that many business sectors are struggling to find workers to fill open positions. The White House even announced a set of new initiatives to lure more Americans back into the workforce.

There are other reasons to be skeptical about grim forecasts regarding AI and employment. First, many past predictions about technologically induced unemployment were wrong because, as noted in a new R Street Institute report on the history of automation fears, we often lack the imagination to describe future jobs or worker skills. A review of old government labor market forecasts or economic papers finds no mention of today’s hottest jobs or skills. Glassdoor’s 2022 best jobs list includes job titles such as: full stack engineer, enterprise architect and machine learning engineer. These jobs would not have been comprehensible to analysts or economists in past decades.

Second, pundits often fail to appreciate how humans adapt rapidly in the face of technological change. Where experts see ominous threats because of new technologies, many others see an opportunity to create new businesses and jobs. A new book, “Working with AI: Real Stories of Human-Machine Collaboration,” offers dozens of case studies of firms integrating algorithmic technologies in the workplace today and shows how organizations are “practicing augmentation, not large-scale automation.” We are using our machines to create entirely new skills and professions.

We’ve seen this story before. Until the 1960s, human “calculators” did hard math on paper and chalkboards until mainframe computers came along and took over those jobs. But that automation freed up those workers to build even better computing machines. The result was the digital revolution, with entrepreneurs and workers seizing new opportunities.

Finally, many technological trends get over-hyped but often fail to materialize. The past decade has seen a lot of hype about autonomous vehicles, leading to much speculation about the potential for job loss for professional drivers. But it turns out that robotic driving is much harder than anticipated and there continues to be a massive shortage of human drivers. 

Market forecasts also tend to overlook how social norms and cultural resistance affect technological adoption. A robot could potentially cut your hair or even be your therapist, but in both cases most people will want an actual human doing that job. Many tasks on airplanes today are handled by autopilot technology, which has dramatically improved aviation safety, but we still want human pilots in the cockpit.

Many jobs will be susceptible to AI and automation, however. Government can try to help reskill workers, even though past retraining programs have not fared well. To better prepare the workforce of the future, policymakers can use a mix of policies to enhance STEM education, tax deductions for retraining, better online learning programs, technical recertification programs, portable benefits solutions and vocational apprenticeship models. It is equally important that lawmakers relax barriers to labor mobility and employment flexibility, especially occupational licensing rules. 

But AI isn’t going away and to prosper, we’ll need to learn how to quickly adapt alongside our latest technological creations.

Adam Thierer is a senior research fellow in technology and innovation at the R Street Institute