Is AI affecting Indiana's labor market?

Workforce Economist, Indiana Department of Workforce Development

Nationally, there has been little movement in the proportion of workers employed in AI-exposed occupations since the release of ChatGPT.

The release of ChatGPT in late 2022 set off a flurry of speculation about how artificial intelligence (AI) will affect labor markets. In early February 2026, 19% of Indiana businesses claimed to be using AI in some way (see Figure 1).

Figure 1: Percent of Indiana businesses using AI in the past two weeks

Note: Prior to November 2025, businesses were asked if they use AI "in producing goods and services." In November 2025, the language of this question was broadened to ask if the business uses AI "in any of its business functions." The change in wording resulted in a break in the comparability of answers before and after.
Source: U.S. Census Bureau, Business Trends and Outlook Survey

But labor market data has been murkier than some might expect. Unemployment rates have risen for recent college graduates compared to the overall workforce, but unemployment has also increased for young people without college degrees. Academic research on the labor market impact of AI is still new and consists largely of unpublished working papers. Some analysts suggest that employment and wages among early-career workers in AI-exposed occupations have been dropping since the advent of generative AI. Other analyses find no effects, or find that workers in AI-exposed occupations are seeing increasing wages. Importantly, the timing of declines in AI-exposed employment and job postings suggests that negative results could reflect factors unrelated to AI.

Defining exposure

For this analysis, several recently developed measures of occupational exposure to AI¹ are combined to define occupational exposure on a scale of 0 to 100. These measures tend to identify the percentage of tasks or work activities that AI can perform. In this analysis, workers or jobs are defined as being exposed to AI if they are in occupations that represent the top 20% of workers exposed to AI (score between 80 and 100). The exposure scores do not directly identify whether AI is likely to augment or replace an occupation’s work.

Figure 2: Occupations by AI exposure score vs. percent of workers with a college degree

Note: Occupations shown had at least 50 respondents.
Source: IPUMS-USA, 2022-2024

The exposed occupations were analyzed further to create several labor market indicators that will be helpful for characterizing AI’s impact on Indiana’s economy. The resulting dashboard will be updated monthly and can be found here.

Employment trends

The first indicator measures the proportion of workers in AI-exposed occupations. If AI were being used mainly to replace labor in exposed occupations, the share of workers in those occupations would be expected to fall.

Nationally, there has been little movement in the proportion of workers employed in exposed occupations since the release of ChatGPT (see Figure 3). There is a long-term trend of technical, exposed occupations increasing and interactive, exposed occupations decreasing. These trends do not appear to have been interrupted by the advent of generative AI.

Figure 3: Percent of U.S. workers in AI-exposed occupations by exposure group

Note: All data are seasonally adjusted.
Source: Current Population Survey (CPS) microdata via IPUMS-CPS

In Indiana, the story appears similar, although Indiana has a lower proportion of workers in technical, exposed occupations (see Figure 4).

Figure 4: Percent of Indiana workers in AI-exposed occupations by exposure group

Note: All data are seasonally adjusted.
Source: CPS microdata via IPUMS-CPS

Unemployment signals

National unemployment rates rose for all workers in 2025, but unemployment among workers in AI-exposed occupations has increased faster, closing the historical gap between the two. In 2022, workers in unexposed occupations had an unemployment rate of 3.4%, compared to just 2.6% for workers in AI-exposed occupations. By November 2025, the rate was 4.1% for unexposed workers and 4.0% for AI-exposed workers (see Figure 5). This closing of the unemployment gap between workers in exposed occupations and those in unexposed occupations might be the first sign of trouble for workers in exposed occupations, but it is probably too early to tie these struggles to AI technologies.

Figure 5: U.S. unemployment rate for AI-exposed and AI-unexposed occupations

Note: All data are seasonally adjusted.
Data: CPS microdata via IPUMS-CPS

Demand for AI-exposed work

Indiana job postings have declined substantially from the abnormally high levels during the post-pandemic recovery. Postings in AI-exposed occupations have decreased by 41% since 2022, while postings in unexposed occupations have decreased by 35% (see Figure 6).

Figure 6: Unique job postings in Indiana by AI exposure type

Source: Lightcast jobs postings

The larger percent decrease among AI-exposed occupations appears consistent with AI-driven displacement in the demand for workers, but as other analyses have pointed out, this decrease seems to have begun prior to the release of ChatGPT. Instead, it might simply reflect a reversal of the post-pandemic hiring surge that was especially strong in tech-related roles.

Advertised wages on job postings in exposed occupations have also increased faster than those in unexposed occupations in Indiana. For AI-exposed jobs, real advertised wages have increased by 16% since 2022, while for AI-unexposed job postings, they have increased by just 11% (see Figure 7). This might suggest that AI plays a complementary role with labor in AI-exposed occupations, particularly in technical, exposed jobs.

Figure 7: Average salary on Indiana job postings by AI exposure type

Source: Lightcast jobs postings

What this means for Indiana

Early evidence suggests that it is possible that AI is having an impact on Indiana’s labor market, but it’s still too early to say so definitively. Past experiences with technological innovations, such as computers and the internet, remind us that it often takes longer than expected for new tools to become incorporated into the economy in a way that is easy to see in aggregate statistics. Even so, rising unemployment and falling job postings among exposed workers might imply some displacement effects, while rising wages in exposed occupations suggest that AI could also be creating opportunities for workers who can effectively use these tools. For Indiana, this preliminary analysis points to a continued need to monitor whether workers are being displaced by new technologies, and to identify and develop skills that will be augmented, rather than replaced, by AI tools.