ECONOMY

Exposure to Generative AI and Expectations About Inequality

With the rise of generative AI (genAI) tools such as ChatGPT, many worry about the tools’ potential displacement effects in the labor market and the implications for income inequality. In supplemental questions to the February 2024 Survey of Consumer Expectations (SCE), we asked a representative sample of U.S. residents about their experience with genAI tools. We find that relatively few people have used genAI, but that those who have used it have a bleaker outlook on its impacts on jobs and future inequality.

Our Sample

Our respondents were 55 percent female, with a median age of 49 years. Among respondents, 4 percent self-identified as American Indian, 4 percent as Asian or Pacific Islander, 9 percent as Black, 8.8 percent as Hispanic, and 84 percent as White. In terms of work, 45 percent of our respondents were working full time, with another 12 percent working part time. More than 39 percent of respondents had a college degree or more education.

Exposure to GenAI Tools

Of our sample, 31 percent had used a genAI tool, as shown in the chart below. Those who had used these tools tended to be younger and were more likely to be male and college educated. They were also less likely to be White.

GenAI Users More Likely to Be Male, Educated, Working and Less Likely to Be White

Source: February 2024 Survey of Consumer Expectations.

Expectations About Impacts of GenAI Tools on Work

Productivity: Among those who had used genAI tools, 60 percent believed the tools had not made them any more or less productive at work, and 35 percent thought that they had enhanced their productivity at work. This result is consistent with the fact that most of the respondents who had used genAI tools used them to obtain information and advice (66 percent) or for entertainment (48 percent), while only 39 percent used them for work. Among those who had used genAI for work, 63 percent believed it had enhanced their productivity.

Wages and employment: In general, a substantial share of respondents did not anticipate that genAI tools would affect wages: 47 percent expected no wage changes. These beliefs did not differ significantly based on prior exposure to genAI tools.

However, respondents believed that genAI tools would reduce the number of jobs available. Forty-three percent of survey respondents overall thought that the tools would diminish jobs. This expectation was slightly more pronounced among those who had used genAI tools, a statistically significant difference.

Respondents were fairly concerned about the impact of genAI tools on their own likelihood of losing a job. Overall, 10 percent of respondents thought they would lose their job on account of these tools. Of course, concern about losing one’s job may be correlated with demographic traits like education, which are also related to having used genAI in the past. After accounting for these other demographic traits, people who had used genAI were 6.5 percentage points more likely to think they would lose their job, a statistically significant difference.

Skills: However, genAI may also be helpful in building skills to retain a job or secure a new one. People who had used genAI tools were more than twice as likely to think that these tools could help them learn new skills that may be useful at work or in locating a new job. Specifically, among those who had not used genAI tools, 23 percent believed that these tools might help them learn new skills, whereas 50 percent of those who had used the tools thought they might be helpful in acquiring useful skills (a highly statistically significant difference, after controlling for demographic traits).

Expectations About Impacts of GenAI Tools on Inequality

We find that those who have used genAI tools tend to be more pessimistic about future inequality. Specifically, we asked people whether they thought there would be more, less, or about the same amount of inequality as there is today for the next generation. The chart below shows that while 33 percent of those who have not used genAI tools think there will be more inequality in the next generation, 53 percent of those who have used genAI tools think there will be more inequality. This gap persists and is statistically significant, even after controlling for other observable traits.

More GenAI Users Expect Inequality to Increase in the Future Than Non-Users

Source: February 2024 Survey of Consumer Expectations.

While suggestive, the descriptive statistics shown in the chart cannot tell us whether people who are grimmer about the future tend to be the ones who have tried out genAI tools, whether exploring the tools makes a person more pessimistic, or whether there are other traits that are associated both with trying genAI and with a bleaker forecast about inequality. More research would be valuable to understand the nexus between genAI use and beliefs about future inequality.

Conclusion

We find that relatively few survey respondents have used genAI tools. Survey respondents who have used these tools for work felt the tools made them more productive, but people who used them otherwise did not think they changed their productivity. Respondents anticipated that these tools would not impact wages but would decrease the number of jobs available. Those who have used genAI tools were more pessimistic about inequality in the future.

Natalia Emanuel is a research economist in Equitable Growth Studies in the Federal Reserve Bank of New York’s Research and Statistics Group.

Emma Harrington is an assistant professor at the University of Virginia.

How to cite this post:
Natalia Emanuel and Emma Harrington, “Exposure to Generative AI and Expectations About Inequality,” Federal Reserve Bank of New York Liberty Street Economics, October 2, 2024, https://libertystreeteconomics.newyorkfed.org/2024/10/exposure-to-generative-ai-and-expectations-about-inequality/.


Disclaimer
The views expressed in this post are those of the author(s) and do not necessarily reflect the position of the Federal Reserve Bank of New York or the Federal Reserve System. Any errors or omissions are the responsibility of the author(s).


Source link

Back to top button