AI is indeed changing the labor market, see the flood of news articles on layoffs happening in part due to companies’ priorities shifting to AI. Now a new working paper from Massachusetts Institute of Technology economist David Autor says that the shift presents a unique opportunity: AI could enable more workers to perform higher-stakes, decision-making tasks that are currently relegated to highly-educated workers such as doctors and lawyers.

Thanks to computers, information is now cheap, abundant, and accessible. As Autor puts it, that’s handed more decision-making power to “elite experts,” or the minority of American adults who hold college or graduate degrees. Computers enabled professionals to spend more time engaging in actual decision-making, rather than just acquiring information. At the same time, these personal machines have also led to the automation of administrative support, clerical, and blue-collar production jobs. As a result, 60% of Americans without a bachelor’s degree have been shifted to non-expert, low-paid service jobs — some of that due to automation — which has contributed to widening economic inequality.

But AI could reverse that shift, Autor finds — and push back against the process started by computerization. With training, new tools can enable workers to accomplish tasks that were previously not possible. Using AI to enable workers to make higher-stakes decisions could broaden access to quality jobs for more workers.

“In essence, AI used well can assist with restoring the middle-skill, middle-class heart of the US labor market that has been hollowed out by automation and globalization,” Autor writes.

Human expertise, a brief history

The economic value of human expertise has changed throughout history. Artisans were first valued for their expertise — in the form of so-called “procedural knowledge,” or following highly-practiced steps to produce an outcome, and “expert judgment,” meaning adapting those procedures to variable instances — in making goods before mass production came along and confined workers to rank-and-file jobs that required fewer specialized skills.

Factory workers toiled under punishing conditions for extremely low pay. The handloom weavers and knitters they replaced were rapidly wiped out by factories after 1815.

As machines gained sophistication, trained workers who could operate and maintain complex equipment were needed. The demand for a new form of worker expertise — “mass expertise” developed.

New jobs such as telephone operators, typists, and bookkeepers, which required literacy and numeracy skills, were formed. At the same time, a growing share of the workforce had a high school diploma. But expert judgment wasn’t needed. Rather, workers were required to follow rules and exercise little discretion, which made the work “perhaps uniquely vulnerable to technological displacement in the era that followed,” Autor writes.

But the rise of personal computers, adopted by businesses in the 1970s altered that dynamic. The computer could cheaply, reliably, and rapidly execute on rules or routine tasks — bringing an end to the need for mass expertise. Enter elite experts, or those who hold four-year degrees. Computers enabled professionals to spend less time acquiring and organizing information and more time interpreting and applying that information, combining procedural knowledge with expert judgment. Equipping elite experts with computers also magnified their value: As computers became more sophisticated, the earnings of workers with college and graduate degrees rose steeply.

And what happened to the workers in the telephone operator and typist jobs? Many shifted into service occupations, placing downward pressure on already low wages for these jobs, and contributing to rising inequality.

AI could improve the quality of jobs for people without college degrees

With AI’s ability to combine rules with acquired experience to support one-off, high-stakes decisions, the technology’s primary role will be to advise, coach, and alert decision-makers as they apply expert judgment, writes Autor.

It’s already happening. Executives tell Quartz that they have seen “significant improvements” in the onboarding process for employees like engineers, accelerating how fast a new hire becomes productive within the company with the internal AI chatbots available at hand.

Instead of devaluing human expertise, AI could support and supplement judgment, enabling more non-elite workers to become a part of high-stakes decision-making processes in their careers. In that way, AI would have an equalizing effect, Autor writes, adding it could “temper the monopoly power that doctors hold over medical care, lawyers over document production, software engineers over computer code, [and] professors over undergraduate education.” Not only could AI improve jobs for workers without college degrees, but it also has the potential to reduce economic inequality and lower the cost of costly services such as healthcare, education, and legal expertise, he argues.

Yes, jobs will be automated, but that’s perhaps not the real problem

The onward march of technology has been automating industries for years. As farming machinery has become more sophisticated, for example, the share of US employment in agriculture has declined from 35% in 1900 to about 1% in 2022.

At the same time, plenty of jobs have been created by technology in that same time. “The majority of contemporary jobs are not remnants of historical occupations that have so far escaped automation,” Autor writes. Instead, they are jobs born from specific technological innovations, demanding new expertise. Consider software engineers, personal trainers, financial influencers, vegan chefs, or college admissions consultants.

In the age of AI, Autor concludes, the challenge for the US and other industrialized nations will be not a shortfall of work, but a shortfall of workers in part due to an aging population.