Over the past three decades, high skill jobs requiring expert problem solving and complex communication skills significantly expanded, with the earnings of the college educated workers needed to fill such jobs rising steadily. Conversely, opportunities and wages declined for less educated middle skill blue-collar manufacturing jobs and white-collar administrative jobs whose careers have been upended by automation.
“But the new kind of automation — artificial intelligence systems called large language models, like ChatGPT and Google’s Bard — is changing that,” noted a recent NY Times article, “In Reversal Because of A.I., Office Jobs Are Now More at Risk.” “These tools can rapidly process and synthesize information and generate new content. The jobs most exposed to automation now are office jobs, those that require more cognitive skills, creativity and high levels of education. The workers affected are likelier to be highly paid, and slightly likelier to be women, a variety of research has found.”
The NYT article referenced two that analyzed the occupations most exposed to AI, one from the Pew Research Center, the other from Goldman Sachs Research.
“Historically, changes in technology have often automated physical tasks, such as those performed on factory floors,” said the Pew report, “Which U.S. Workers Are More Exposed to AI on Their Jobs?.” “But AI performs more like human brainpower and, as its reach grows, that has raised questions about its impact on professional and other office jobs – questions that Pew Research Center seeks to address in a new analysis of government data.” The Pew analysis is based on data on 41 essential work activities in 873 occupations from the U.S. Department of Labor’s Occupational Information Network (O*NET)
Let me summarize the Pew report’s key findings.
- In 2022, about 20% of American workers were in jobs with high exposure to AI, — that is, their most important activities may be either replaced or assisted by AI, — while 23% of workers were in jobs in which their most important activities are least exposed to AI.
- Jobs with high exposure to AI include budget analysis, tax preparers, technical writers, and Web developers; medium exposure jobs include chief executives, veterinarians, interior designers, and sales managers; and low exposure jobs include child care workers, dishwashers, barbers, and firefighters.
- Jobs with a high level of exposure to AI tend to be in higher-paying fields; workers in the most exposed jobs earned $33 per hour, on average, compared with $20 in jobs with the least amount of exposure.
- Workers with a bachelor’s degree or more (27%) are more than twice as likely to face high exposure to AI as those with only a high school diploma (12%).
- Most workers are likely to be in jobs with less exposure to AI than in jobs with more exposure, especially men, younger workers, those with less formal education, and Black and Hispanic workers.
- A greater share of women (21%) than men (17%) are likely to see the most exposure to AI, because of differences in the types of jobs held by men and women.
- U.S. workers in more exposed industries do not feel their jobs are at risk; for example, about one third of workers in information and technology say AI will help more than hurt them personally, compared with 11% who say it will hurt more than it help.
Let’s now turn to the Goldman Sachs report, “The Potentially Large Effects of Artificial Intelligence on Economic Growth.” “Despite significant uncertainty around the potential of generative AI, its ability to generate content that is indistinguishable from human-created output and to break down communication barriers between humans and machines reflects a major advancement with potentially large macroeconomic effects,” said the report. “Generative AI’s ability to 1) generate new content that is indistinguishable from human-created output and 2) break down communication barriers between humans and machines reflects a major advancement with potentially large macroeconomic effects.”
To assess the share of total work exposed to AI-based automation by occupation and industry, the Goldman Sachs report analyzed data on over 900 US occupations from the O*NET database, as well as data on over 2000 occupations from the European Commission’s ESCO database.
“If generative AI delivers on its promised capabilities, the labor market could face significant disruption. Using data on occupational tasks in both the US and Europe, we find that roughly two-thirds of current jobs are exposed to some degree of AI automation, and that generative AI could substitute up to one-fourth of current work.” Extrapolating its estimates globally suggests that generative AI could have an impact on 18% of jobs around the world, the equivalent of 300 million full-time jobs.
Let me summarize some of the report’s key findings.
- Occupations whose current work tasks are most exposed to AI automation in the US are: office and administrative support (46% of work tasks); legal (44%); architecture and engineering (37%); biological, physical, and social science (36%); business and financial operations (35); community and social service (33%); and management (32%).
- Occupations whose current work tasks are least exposed to AI automation in the US are: building and ground cleaning and maintenance (1%); installation, maintenance; and repair (4%), construction and extraction (6%); production (9%); transportation and material moving (11%); food preparation and serving (12%); and personal care and service (19%).
- Occupations whose current work tasks are most exposed to AI automation in Europe are: clerical support (45% of work tasks); professionals (34%); technical and associate professionals (31%); and managers (29%).
- Occupations whose current work tasks are least exposed to AI automation in Europe are: craft and related trades (4%); plant and machine operations (7%); elementary physical, routine occupations (8%); service and sales workers (15%); skilled agricultural, forestry, and fisheries (21%); and armed forces occupations (22%).
- Most jobs and industries are only partially exposed to automation and are thus more likely to be complemented rather than substituted by AI: 63% of current US employment are expected to be complemented, 7% are expected to be substituted, and 30% are expected to be unaffected.
- Overall, AI is expected to enhance the productivity of the vast majority of occupations; the largest impact of AI automation is expected in legal and administrative fields; and the least impact is expected in manual and outdoor jobs.
Generative AI has the potential for a boom in labor productivity and a significant increase in global output. “[M]ost workers are employed in occupations that are partially-exposed to AI automation and, following AI adoption, will likely apply at least some of their freed-up capacity toward productive activities that increase output.”
In addition, “many workers that are displaced by AI automation will eventually become reemployed — and therefore boost total output — in new occupations that emerge either directly from AI adoption or in response to the higher level of aggregate and labor demand” generated by the AI productivity boost.
The timing of an AI labor productivity boom is hard to predict, but, based on the history of previous transformative technologies, the boom generally starts about two decades after the technological breakthrough, when roughly half of US businesses have adopted the technology. But the AI labor productivity boom could happen faster because much of the necessary infrastructure to acquire and start using new digital technologies is already in place, including the internet, cloud computing, software-as-a-service, app stores, and other advances. “Our key takeaway,” said the Goldman Sachs report in conclusion, “is that the ultimate boost to labor productivity is uncertain, but in most scenarios would remain economically significant.”
The NY Times article further discussed the implications of generative AI mostly complementing rather than replacing around two thirds of US occupations. All of a sudden, many workers will have an AI assistant helping them learn new skills and advance faster in their careers. The article references two recent working papers that found that generative AI is particularly helpful to junior employees.
Generative AI at Work, a working paper by Stanford professor Erik Brynjolfsson and his collaborators found that using a GenAI-based tool increased the overall productivity of customer support agents by 14% on average. The greatest impact, around 35%, was on novice and low-skilled workers, while the impact on experienced and highly skilled workers was minimal.
The second working paper, “Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence,” was based on a randomized trial of 444 college-educated, mid-level professionals to assess the productivity impact of generative AI on writing tasks in fields like human relations and marketing. Its results showed that ChatGPT substantially raised the average productivity of those who used the technology by a substantial 37%. In addition, the use of ChatGPT reduced inequality by benefiting low-skill workers more than high-skill ones.
“The last round of automation, affecting manufacturing jobs, increased income inequality by depriving workers without college educations of high-paying jobs, research has shown,” said the NYT article in conclusion. “A.I. could perhaps do this again — for example, if senior managers called on large language models to do the work of junior staffers, potentially increasing the earnings of executives while displacing the jobs of those with less experience.” But generative AI may well end up reducing the inequality between the highest-paid workers and everyone else.
Comments