Hospital Market Concentration and Hospital Adoption of Artificial Intelligence and Machine Learning

Abstract Background We examined associations of hospital market concentration and organizational characteristics with artificial intelligence (AI) and machine learning (ML) use. Methods We analyzed 2025 hospital survey data, measuring concentration with the Herfindahl–Hirschman Index (HHI). Primary outcomes measured current use of AI/ML both integrated into and outside electronic health records (EHRs), and function counts requiring tools from both the EHR developer and another source. Alternative measures allowed either setting or source. Standard errors were clustered by hospital referral region. Results High versus low concentration was associated with a 6.7-percentage-point higher adjusted probability of generative-AI use in both settings (95% CI, 0.2–13.2), but not any use (0.2; 95% CI, −4.9 to 5.2). Moderate concentration was associated with 0.20 fewer generative functions requiring both sources (95% CI, −0.39 to −0.01). System affiliation and accountable care organization (ACO) participation were positively associated with all five outcomes under both definitions. Conclusions Market concentration was not consistently associated with AI/ML use across technologies and outcome definitions. System affiliation and ACO participation were consistently associated with greater use.