A More Nuanced Analysis of AI's Effect on Jobs

An insightful new paper from OpenAI’s research group begins:

Most analysis of AI’s impact on the labor market begins with the same core question: what jobs are most exposed to AI? That is an important starting point, but it is not enough when it comes to measuring the risk of major labor market disruption. Exposure helps us understand where AI has technical capability. It cannot, on its own, tell us which jobs are most likely to be automated, redesigned, or expanded in the near term.

The paper introduces “the AI Jobs Transition Framework” that attempts to shift the conversation away from analyzing which occupations are most exposed to AI and toward analyzing three factors: the degree to which an occupation requires a human worker, the relationship between demand for goods and services and the price of those goods and services, and the degree to which an occupation is technically exposed to AI.

The third is familiar, but let’s take the first two in turn.

Human beings will still need to be teachers in classrooms, even if generative AI tools do a lot of the tasks currently performed by teachers. It may be that one day generative AI could do the work of federal judges, but it is unlikely that the legal system would allow a large language model to issue criminal sentences. Some occupations require a service provider to engender trust and persuasion that AI tools are unlikely to provide (though that may change). These sorts of occupations require human workers to a larger degree than, say, customer service representatives.

Generative AI tools will increase productivity, allowing firms to sell goods and services for lower prices. This could reduce employment in some occupations. But it could raise employment in other occupations, if the price cut induces a relatively large increase in market demand for the good or service being sold by a given firm, which in turn would increase that business’s need for workers. The degree to which AI adoption affects employment in specific occupations depends in part on how sensitive demand is to the price of the output those workers are helping to produce. For example, AI-powered reductions in the price of software could increase employment of software engineers. But cheaper prices are likely to have a much smaller effect on the employment of CEOs, since each company needs only one CEO.

https://x.com/MichaelRStrain/status/2045236189582733762

OpenAI economists combined those three factors to create a more nuanced picture of the near-term impact of AI on employment. They find that 18 percent of jobs — nearly 1 in 5 — have a high near-term risk from AI. They further find that 12 percent of occupations could see employment increases due to AI. They expect around one-quarter of jobs to reorganize and 46 percent of jobs not to face near-term change.

I have been working with OpenAI and am quite impressed by their commitment to analyze the information they are getting from their users in order to identify how AI tools are changing the economy. Their commitment to disseminating that information to decision-makers and into the public debate is equally impressive.

Check out the full paper here, which goes into much more detail and presents additional analysis.

This analysis jibes with the approach to technological change increasingly taken by economists: think of a job as a bundle of tasks and imagine how a new technology might change those tasks. An example from my recent National Affairs article on AI optimism:

Retail-store managers — a third example — will need to spend less time managing employees' schedules and the cycle of inventory; AI tools will be able to complete those tasks for them. This will give managers more time to oversee and coach workers, solve problems, and create a positive shopping experience for customers. AI will also assist managers by making suggestions to optimize the shopping experience in the store and proposing potential management strategies based on an employee's career history and other factors.

Of course, over the longer term, we know very little. Again, from my National Affairs article:

The AI revolution will create many opportunities that we cannot conceive today. Standing in the year 2024 and trying to predict the jobs of the future is no easier than standing in the year 1944 and trying to predict that the labor market of the future would contain systems analysts, circuit-layout designers, fiber scientists, and social-media managers. In fact, about 60% of jobs held by workers in 2018 had not been invented as of 1940. New occupations emerge in large part because technology advances, creating new goods and services that in turn require human workers to engage in new occupational tasks. Technological advances also make society wealthier, increasing the demand for goods and services — especially new goods and services — which in turn raises the demand for workers' skills, talents, and efforts.

To illustrate, imagine trying to explain to the 19th-century classical economist David Ricardo the jobs of all the people who support Bruce Springsteen's records and tours: sound engineers, digital editors, graphic designers, photographers, videographers, art directors, instrument technicians, social-media directors, marketing professionals, bookers, stage hands, sound directors, lighting engineers, body men, commercial-vehicle drivers, and, of course, the jobs of the members of the mighty E Street Band and Mr. Springsteen himself. These occupations and the tasks workers perform for them did not exist in Ricardo's time because the technology that enables them had not been invented. They also did not exist because the wealth created by today's technology had not been generated: Society in Ricardo's day could not have afforded rock bands.

Although we cannot make specific predictions of the economic effects of AI with any accuracy, evidence from the past should allow us to make two general predictions with a high degree of confidence: that AI technology will be disruptive, creating winners and losers in the labor market in part by creating entirely new occupations; and that the net effect of AI advances will be to increase workers' productivity, wages, and incomes, which will benefit them overall.