Study Finds U.S. AI Adoption Uneven, Concentrated in Large Companies and Specific Industries
A new working paper from the National Bureau of Economic Research reveals that artificial intelligence (AI) adoption in the United States is occurring unevenly, with significant disparities among companies and industries. The study, which Kristina McElheran, a visiting scholar with the MIT Initiative on the Digital Economy, is leading, offers a nuanced look at the early adoption of AI and challenges the prevalent narrative of universal integration.
Key findings from the study include:
Low Overall Adoption: In 2017, only 6% of U.S. companies reported using AI for production. A November 2023 Census Bureau survey showing that less than 4% of businesses use AI to produce goods and services demonstrates how low this number has remained.
Large Companies Lead Adoption: AI adoption is concentrated among more prominent corporations. More than 50% of companies with over 5,000 employees and over 60% of those with more than 10,000 employees are utilising AI.
Sector Variability: Adoption rates vary across industries, with approximately 12% of firms in manufacturing, information services, and healthcare incorporating AI. In contrast, only 4% of companies in construction and retail reported AI usage.
Regional Clustering: While AI adoption is prominent in "superstar" cities like San Francisco, San Antonio, and Nashville, the study highlights unexpected places, such as manufacturing hubs in the Midwest and certain Southern cities, where AI adoption is also clustered.
Startups and Leadership Characteristics: Startups embracing AI tend to have younger, highly educated, and experienced leaders. Additionally, venture capital backing and a focus on process innovation are associated with higher AI adoption rates.
McElheran emphasises that the data challenges the prevailing notion that AI is rapidly permeating all sectors simultaneously. The study suggests that AI integration is complex and varies based on company size, industry, and leadership characteristics.
The study acknowledges the potential dark side of AI adoption, citing inertia and adjustment costs as significant barriers. While AI promises benefits, including increased efficiency and productivity, it poses challenges, such as potential job losses and economic adjustments. The findings call for a realistic and evidence-based approach to AI adoption, considering its benefits and costs across different segments of the economy and society.