A new poll reveals what Americans fear about AI taking their jobs

The explosive launch of OpenAI’s ChatGPT in November 2022 has put the artificial intelligence or AI conversation into overdrive. As we sort through potential ramifications of advanced AI applications, the impact on jobs is at the top of the list.

In July, two significant reports on AI job displacement were released. On July 11, the Organisation for Economic Co-operation and Development reported, “27% of jobs are in occupations at high-risk of automation.”

Later that month, Pew Research reported a similar finding: “In 2022, 19% of American workers were in jobs that are the most exposed to AI, in which the most important activities may be either replaced or assisted by AI.”

The explosive launch of OpenAI’s ChatGPT in November 2022 has put the artificial intelligence or AI conversation into overdrive. As we sort through potential ramifications of advanced AI applications, the impact on jobs is at the top of the list.

In July, two significant reports on AI job displacement were released. On July 11, the Organisation for Economic Co-operation and Development reported, “27% of jobs are in occupations at high-risk of automation.”

Later that month, Pew Research reported a similar finding: “In 2022, 19% of American workers were in jobs that are the most exposed to AI, in which the most important activities may be either replaced or assisted by AI.”

Big picture, our poll revealed that Americans understand the importance of AI: half of Americans (48%) believe it to be important to America’s future economic success.

When it comes to potential job displacement by AI, it turns out that Americans feel confident AI won’t take their job but are worried it will take other people’s jobs.

First we asked if Americans were worried about AI displacing Americans’ jobs “generally.” The options were “very worried,” “somewhat worried,” and “not worried at all.” Only one in five Americans aren’t worried at all. Most Americans are worried: four out of five were at least somewhat worried to very worried. A little more than one out of five Americans are very worried.

Next we asked Americans if they were worried about AI displacing their own job. Stunningly, three out of five Americans were “not worried at all.” Only one in 10 Americans were “very worried.” This poll was conducted with a representative sample of the whole American population.

In fact, this finding matches closely with a survey conducted as part of a separate Center for Growth and Opportunity research project on the future of work. This research also used a survey to gauge Americans’ attitudes toward technology and its impact on work and specifically asked if Americans thought their job would exist 30 years from now. Three out of four Americans answered saying their job would exist in 30 years.

Automation and robotics, which can sometimes qualify as AI but not always, have long been known to displace workers in repetitive tasks like automobile assembly. However, it’s not all bad news. My coauthor Erica Jednyak and I pointed out in 2020 that these displacements often mean lower costs and higher quality goods for Americans.

ChatGPT and generative AI have exposed a whole new category of creative and intellectual jobs to potential displacement that more repetitive tasks have been exposed to. In the Pew research cited above, the top three job categories with the most “exposure” to AI fall into these categories: professional, scientific and technical services; finance, insurance, and real estate; and public administration.

This is why we asked respondents for their thoughts on whether AI would “help” or “hurt” those employed in specific industries including law, medical, engineering, journalism, customer service, academia, publishing, management, K-12 education, manufacturing and transportation.

Out of all these job categories, Americans thought AI was most likely to hurt journalism and customer service by 10 points compared to other jobs.

All of these numbers tell us a few things about how Americans are approaching this new technology.

The fact that most Americans aren’t worried about AI taking their own job away perhaps reflects the intimate knowledge they have for their own jobs. When a discussion about “jobs out there” or “generally,” as we described it, that can be an amorphous concept that misses all the complexities that go into even simple jobs.

For example, long-haul trucking has long been thought to be threatened by autonomous trucks operated by AI, yet companies and researchers continue to discover new complexities on the road that require human input. Full replacement has remained elusive and likely will for many years.

This trend matches with the historical adoption of new technologies. Fears of job displacement by new technologies are woven through human history. However, history also demonstrates that while some jobs are lost, even more jobs are created or current jobs are made more productive. Americans seem to get that.

This finding could also perhaps show a blindspot in awareness of the capabilities of AI. All workers in all fields should be cognizant of the potential uses and applications in their field. As history has demonstrated so far with AI tools, outright displacement is not likely, rather augmentation and integration of AI into current jobs is the more likely outcome.

With generative AI tools like ChatGPT, it is no longer repetitive predictable tasks that could be displaced or changed, but imaginative or creative tasks like writing and analysis.

Our poll can inform policymakers and journalists covering AI that the American people have nuanced views on this exciting technology. Those two groups can use their platform to provide sober analysis and information as we all continue to learn about the ramifications of AI in our economy.

CGO scholars and fellows frequently comment on a variety of topics for the popular press. The views expressed therein are those of the authors and do not necessarily reflect the views of the Center for Growth and Opportunity or the views of Utah State University.