Has Generative Artificial Intelligence Adoption Impacted Labor Demand at Third District Firms

Introduction

As generative artificial intelligence (AI) tools have become more widely accessible over the past few years, researchers have sought to understand how the technology will change the future of work. Generative AI diverges from traditional AI in its ability to create original content based on vast amounts of data.¹ Familiar examples of generative AI include large language models (LLMs) such as ChatGPT or Google's Gemini, which can process and produce new content, including text or images. These LLMs can write emails, draft computer code, analyze data, and perform other tasks traditionally done by humans.

The authors discuss why studying the adoption of generative
AI at firms within the Third District is important and who may
benefit from this research.

Our recent report found that jobs requiring more education and paying higher salaries are more likely to be affected by AI, such that its usage can reduce the completion time for a higher share of those jobs' required tasks. However, the labor market impacts of this technology will depend on both occupational exposure and the level and speed of AI adoption by firms. The degree to which the technology will lead to automation and worker displacement is a major concern among workers and policymakers. Over 20 percent of workers have used AI in their job, based on findings from a Pew Research Survey conducted in September 2025, about 5 percentage points higher than a year earlier.² According to another nationwide survey conducted by the National Opinion Research Center (NORC) at the University of Chicago and analyzed by the Brookings Institution, about 45 percent of respondents think AI will decrease the number of jobs in their industry, whereas only 11 percent think it will lead to increased job opportunities.³

Results from our recent survey of firms in the Third Federal Reserve District - Delaware, southern New Jersey, and eastern and central Pennsylvania - indicate that half of firms are currently using generative AI tools in the workplace, but few AI-adopting firms reported replacing workers.

Half of Third District Firms Surveyed Report Using Generative AI

Results from the survey⁴ show that nearly three-quarters of businesses that responded currently use some form of AI, as seen in Figure 1. About half of the firms said they use generative AI, while 23 percent use traditional AI or do not know which type they use. The rate of generative AI adoption by firms in our sample is similar to the national rates reported by the Ramp AI Index in October 2025, which provides data on firms subscribed to any AI product or service.

Although most firms reported using some form of AI, about one-quarter of businesses are not using any AI technologies. Just 5 percent of firms do not use AI but plan to adopt it in the next year, while 21 percent of firms have no plans to do so or do not know.

A greater share of service-providing firms reported using generative AI compared with goods-producing firms (see Figure 2a). Over 60 percent of service-providing firms said they currently use generative AI, versus one-third of goods-producing firms. While goods-producing firms were more likely than service-providing firms to use traditional AI alone or not know which type they used, they were also more likely to report not using any form of AI. Twenty-eight percent of goods-producing firms reported not using AI, with no plans to do so, compared with 16 percent of service-providing firms.

The share of firms that adopted generative AI was similar across firm sizes, although it was slightly higher for larger firms. About 57 percent of large firms, 46 percent of midsize firms, and 51 percent of small firms reported using generative AI. At the same time, a greater share of small firms reported not using any AI, without plans for or knowledge of future adoption.

Majority of Firms Reported No Impact on Need for Workers Due to Generative AI

Among the 50 percent of firms that reported using generative AI, the majority did not see an impact on their need for workers. Almost 70 percent of the firms that use generative AI in our sample reported that the adoption of this technology did not affect their need for workers, while 17 percent said that using generative AI changed the type of workers needed but not the number. While more firms said that generative AI decreased their need for workers than increased their need, the share of firms reporting impacts to employee headcount remain relatively small. Eight percent of firms reported that generative AI decreased their need for workers, and 2 percent noted it increased their need for workers. These results are consistent with a 2025 analysis from the Dallas Fed, which indicated that generative AI has had a relatively small impact on demand for workers thus far.⁵

Closing Thoughts

Our results suggest a limited number of firms have reported a decreased need for workers due to their adoption of generative AI. Although AI usage among firms is still in its early stages, these results suggest that many firms may use generative AI to enhance worker skills and improve productivity, rather than automate roles. Findings from a New York Fed study indicate that 47 percent of the service and manufacturing firms surveyed expected to retrain their current workforce to use AI over the next six months, whereas 13 percent of service firms and no manufacturers anticipated layoffs due to AI.⁶

However, recent research suggests that certain types of workers have been more susceptible to job replacement by the technology, such as early career workers in more AI-exposed occupations (e.g., software development, customer support).⁷ To inform workforce development practices and training strategies, future research could explore the types (e.g., administrative, information technology, management) of workers that firms need more or less of after adopting generative AI.